ByteDance
Top Seed Talent Program
Like "seeds of fire", together we will light up the path to the future.
ByteDance
Top Seed Talent Program
Top Seed Talent Program
Like "seeds of fire", together we will light up the path to the future.
Why do we launch Top Seed?
What kind of individuals are we looking for?
What kind of resources and support do we provide?
*01
Top Seed - About the Program
What is Top Seed
The Top Seed Talent Program is an exclusive initiative launched by the ByteDance Doubao Team to attract the brightest minds on campus. We have established dedicated recruitment initiatives for recent doctoral graduates and research internships targeting exceptional current students, aiming to attract and onboard more top talents worldwide who aspire to "change the world with technology".
This program is part of our efforts to enhance investment in top talent and cutting-edge technologies. Our ambition is to discover and cultivate young talents with a forward-looking vision. At ByteDance, you have the opportunity to work with the brightest scientists and engineers to overcome industry's toughest challenges. Like "seeds of fire", together we will light the path to the future.
What We Are Looking For
PhD candidates graduating between September 2024 and August 2025
We hope you are:
/Innovators who are ambitious about changing the world with technology
/Tech enthusiasts with keen interests in areas such as machine learning, artificial intelligence, LLM, computer vision, and audio and video generation
/People who have a proven track record in related academic research, engineering, and open-source communities
/Tech enthusiasts with keen interests in areas such as machine learning, artificial intelligence, LLM, computer vision, and audio and video generation
/People who have a proven track record in related academic research, engineering, and open-source communities
PhD candidates graduating in or after September 2025
We hope you are:
/Passionate about leveraging technology to change the world, aspiring to engage in fields such as machine learning, artificial intelligence, LLM, computer vision, and audio and video generation.
/Essentially excellent, daring to innovate, and highly self-driven. We value the inherent brilliance in you more than your experience and previous achievements.
/Essentially excellent, daring to innovate, and highly self-driven. We value the inherent brilliance in you more than your experience and previous achievements.
Graduate Opportunities
Internship Opportunities
What We Are Looking For
Graduate Opportunities
招聘对象：
2024年9月-2025年8月毕业的博士同学
我们希望你是：
怀有科技改变世界的远大抱负，敢于创新；
有志于投身机器学习、人工智能、大模型、计算机视觉、音视频生成等技术领域；
在相关学术研究、工程实践、开源社区等表现出色；
有志于投身机器学习、人工智能、大模型、计算机视觉、音视频生成等技术领域；
在相关学术研究、工程实践、开源社区等表现出色；
Internship Opportunities
招聘对象：
2024年9月-2025年8月毕业的博士同学
我们希望你是：
怀有科技改变世界的远大抱负，敢于创新；
有志于投身机器学习、人工智能、大模型、计算机视觉、音视频生成等技术领域；
在相关学术研究、工程实践、开源社区等表现出色；
有志于投身机器学习、人工智能、大模型、计算机视觉、音视频生成等技术领域；
在相关学术研究、工程实践、开源社区等表现出色；
*02
Delve deep into trailblazing research and explore diverse application scenarios
Tackle challenges in large language models, natural language processing, computer vision, speech recognition and generation, and machine learning.
LLMs
/ Generalization of reward models and reinforcement learning models
/ Self-learning of reward and reinforcement learning models
/ Interpretability of large language models
/ Authenticity of large language models
/ Next-generation reinforcement learning algorithms
/ Large language models based on self-play
Machine Learning Algorithms and Systems
/ Develop efficient LLM architectures to optimize performance while minimizing training and inference costs.
/ Research massive training clusters to enhance training stability and Machine Fraction Utilization (MFU), facilitating effective cross-cluster training.
/ Address memory-bound issues during inference, investigate multi-machine inference, and develop parallel inference strategies.
/ Integrate next-generation computing systems to advance model architectures, training methods, and inference techniques.
/ Explore algorithm innovation for LLM foundation models.
Multi-Modal Understanding and Generation
/ Develop foundational models for multi-modal understanding and generation across images, audio, and video, pursuing a unified modeling approach.
/ Design and optimize the architecture of multi-modal models and diffusion models, emphasizing efficient large-scale distributed training and inference systems.
/ Explore efficient methods for representing 3D objects and scenes, learn world knowledge from video data, and construct models of the physical world.
/ Unified foundation models for audio understanding and generation, including speech recognition, audio synthesis, voice conversion, music generation, and sound effects.
*03
Power up your growth with our best-in-class mentor team and dedicated professionals
Liang Xiang
Liang Xiang graduated from the Department of Automation, University of Science and Technology of China in 2006. He was directly admitted to the Institute of Automation, Chinese Academy of Sciences, where he received his PhD in machine learning and recommendation systems. In 2009, he participated in the Netflix Prize competition and won second place. He is also the author of Recommendation System Practice and founded Resys China, a recommendation system community in 2009.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
Mingxuan Wang
Mingxuan earned a PhD from the Institute of Computing Technology, Chinese Academy of Sciences. He currently heads the Large Language Model Research team at ByteDance Doubao. In the field of machine translation, he has authored more than 50 papers presented at leading conferences and has consistently achieved first place in the Workshop on Machine Translation (WMT) on several occasions.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Lin Yan
Lin received a master's degree from the Institute of Computing Technology, Chinese Academy of Sciences. Currently, he leads the Large Language Model Post-Training Team at ByteDance Doubao.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
Yuxuan Wang
Yuxuan obtained his PhD from The Ohio State University, specializing in deep learning, speech technology, and other related fields of research and application. He has led several groundbreaking efforts in the speech industry, with results widely applied in academia and the industrial world.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Jiashi Feng
Jiashi holds a bachelor's degree from the University of Science and Technology of China and a doctoral degree from the National University of Singapore. His expertise lies in computer vision and machine learning, as well as their application in multimedia.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Zhi Tian
Zhi earned his PhD in Computer Science from the University of Adelaide, Australia, where he was mentored by Professor Chunhua Shen. In 2019, he was honored with the Google PhD Fellowship.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhuo Chen
Zhuo obtained his PhD from Columbia University in 2017 and worked as a core application scientist at Microsoft after graduation.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
Chenggang Li
Chenggang obtained his bachelor's and master's degrees from the Department of Mechanical Engineering at Tsinghua University.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Liang Xiang
Liang Xiang graduated from the Department of Automation, University of Science and Technology of China in 2006. He was directly admitted to the Institute of Automation, Chinese Academy of Sciences, where he received his PhD in machine learning and recommendation systems. In 2009, he participated in the Netflix Prize competition and won second place. He is also the author of Recommendation System Practice and founded Resys China, a recommendation system community in 2009.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
Mingxuan Wang
Mingxuan earned a PhD from the Institute of Computing Technology, Chinese Academy of Sciences. He currently heads the Large Language Model Research team at ByteDance Doubao. In the field of machine translation, he has authored more than 50 papers presented at leading conferences and has consistently achieved first place in the Workshop on Machine Translation (WMT) on several occasions.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Lin Yan
Lin received a master's degree from the Institute of Computing Technology, Chinese Academy of Sciences. Currently, he leads the Large Language Model Post-Training Team at ByteDance Doubao.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
Yuxuan Wang
Yuxuan obtained his PhD from The Ohio State University, specializing in deep learning, speech technology, and other related fields of research and application. He has led several groundbreaking efforts in the speech industry, with results widely applied in academia and the industrial world.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Jiashi Feng
Jiashi holds a bachelor's degree from the University of Science and Technology of China and a doctoral degree from the National University of Singapore. His expertise lies in computer vision and machine learning, as well as their application in multimedia.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Zhi Tian
Zhi earned his PhD in Computer Science from the University of Adelaide, Australia, where he was mentored by Professor Chunhua Shen. In 2019, he was honored with the Google PhD Fellowship.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhuo Chen
Zhuo obtained his PhD from Columbia University in 2017 and worked as a core application scientist at Microsoft after graduation.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
Chenggang Li
Chenggang obtained his bachelor's and master's degrees from the Department of Mechanical Engineering at Tsinghua University.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Liang Xiang
Liang Xiang graduated from the Department of Automation, University of Science and Technology of China in 2006. He was directly admitted to the Institute of Automation, Chinese Academy of Sciences, where he received his PhD in machine learning and recommendation systems. In 2009, he participated in the Netflix Prize competition and won second place. He is also the author of Recommendation System Practice and founded Resys China, a recommendation system community in 2009.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
Mingxuan Wang
Mingxuan earned a PhD from the Institute of Computing Technology, Chinese Academy of Sciences. He currently heads the Large Language Model Research team at ByteDance Doubao. In the field of machine translation, he has authored more than 50 papers presented at leading conferences and has consistently achieved first place in the Workshop on Machine Translation (WMT) on several occasions.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Lin Yan
Lin received a master's degree from the Institute of Computing Technology, Chinese Academy of Sciences. Currently, he leads the Large Language Model Post-Training Team at ByteDance Doubao.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
Yuxuan Wang
Yuxuan obtained his PhD from The Ohio State University, specializing in deep learning, speech technology, and other related fields of research and application. He has led several groundbreaking efforts in the speech industry, with results widely applied in academia and the industrial world.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Jiashi Feng
Jiashi holds a bachelor's degree from the University of Science and Technology of China and a doctoral degree from the National University of Singapore. His expertise lies in computer vision and machine learning, as well as their application in multimedia.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Zhi Tian
Zhi earned his PhD in Computer Science from the University of Adelaide, Australia, where he was mentored by Professor Chunhua Shen. In 2019, he was honored with the Google PhD Fellowship.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhuo Chen
Zhuo obtained his PhD from Columbia University in 2017 and worked as a core application scientist at Microsoft after graduation.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
Chenggang Li
Chenggang obtained his bachelor's and master's degrees from the Department of Mechanical Engineering at Tsinghua University.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Expand
Liang Xiang
Liang Xiang graduated from the Department of Automation, University of Science and Technology of China in 2006. He was directly admitted to the Institute of Automation, Chinese Academy of Sciences, where he received his PhD in machine learning and recommendation systems. In 2009, he participated in the Netflix Prize competition and won second place. He is also the author of Recommendation System Practice and founded Resys China, a recommendation system community in 2009.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
Expand
Mingxuan Wang
Mingxuan earned a PhD from the Institute of Computing Technology, Chinese Academy of Sciences. He currently heads the Large Language Model Research team at ByteDance Doubao. In the field of machine translation, he has authored more than 50 papers presented at leading conferences and has consistently achieved first place in the Workshop on Machine Translation (WMT) on several occasions.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Expand
Lin Yan
Lin received a master's degree from the Institute of Computing Technology, Chinese Academy of Sciences. Currently, he leads the Large Language Model Post-Training Team at ByteDance Doubao.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
Expand
Yuxuan Wang
Yuxuan obtained his PhD from The Ohio State University, specializing in deep learning, speech technology, and other related fields of research and application. He has led several groundbreaking efforts in the speech industry, with results widely applied in academia and the industrial world.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Expand
Jiashi Feng
Jiashi holds a bachelor's degree from the University of Science and Technology of China and a doctoral degree from the National University of Singapore. His expertise lies in computer vision and machine learning, as well as their application in multimedia.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Expand
Zhi Tian
Zhi earned his PhD in Computer Science from the University of Adelaide, Australia, where he was mentored by Professor Chunhua Shen. In 2019, he was honored with the Google PhD Fellowship.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Expand
Zhuo Chen
Zhuo obtained his PhD from Columbia University in 2017 and worked as a core application scientist at Microsoft after graduation.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
Expand
Chenggang Li
Chenggang obtained his bachelor's and master's degrees from the Department of Mechanical Engineering at Tsinghua University.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
Expand
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Expand
Liang Xiang
Liang Xiang graduated from the Department of Automation, University of Science and Technology of China in 2006. He was directly admitted to the Institute of Automation, Chinese Academy of Sciences, where he received his PhD in machine learning and recommendation systems. In 2009, he participated in the Netflix Prize competition and won second place. He is also the author of Recommendation System Practice and founded Resys China, a recommendation system community in 2009.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
After joining ByteDance, he was a research scientist for video understanding in AILab before taking charge of the recommendation system. Liang joined AML in 2021 and is been at the helm of ByteDance's AML and Doubao Foundation teams now. He leads the charge in exploring foundational and advanced AI algorithms and engineering technologies, swiftly bringing a range of AI products from concept to market.
Expand
Mingxuan Wang
Mingxuan earned a PhD from the Institute of Computing Technology, Chinese Academy of Sciences. He currently heads the Large Language Model Research team at ByteDance Doubao. In the field of machine translation, he has authored more than 50 papers presented at leading conferences and has consistently achieved first place in the Workshop on Machine Translation (WMT) on several occasions.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Mingxuan has participated in multiple open-source projects, including lightseq, mrasp, and others, which are widely adopted in the industry. He also serves as the regional chairman and sponsorship chairman for conferences such as NeurIPS, ACL, EMNLP, and more.
Expand
Lin Yan
Lin received a master's degree from the Institute of Computing Technology, Chinese Academy of Sciences. Currently, he leads the Large Language Model Post-Training Team at ByteDance Doubao.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
The Post-Training team's work includes instruction tuning, reward models, RLHF, RLAIF, and self-learning models. Meanwhile, the team is also spearheading a series of avant-garde and challenging studies in key areas such as the generalizability, interpretability, and authenticity of large language models.
Expand
Yuxuan Wang
Yuxuan obtained his PhD from The Ohio State University, specializing in deep learning, speech technology, and other related fields of research and application. He has led several groundbreaking efforts in the speech industry, with results widely applied in academia and the industrial world.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Currently, he heads the Speech team at ByteDance Doubao, leading the team in core research and applications of technologies focused on multimodal generation and understanding.
Expand
Jiashi Feng
Jiashi holds a bachelor's degree from the University of Science and Technology of China and a doctoral degree from the National University of Singapore. His expertise lies in computer vision and machine learning, as well as their application in multimedia.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Jiashi currently heads the Vision Research team at ByteDance Doubao, conducting cutting-edge research in visual and multimodal foundation models, AIGC, and 3D avatar/object reconstruction and generation, while also focusing on the wide application of these technologies. Before joining ByteDance, he was an assistant professor in the ECE department of the National University of Singapore, leading the Vision and Machine Learning Laboratory, and advising or co-advising nearly 20 doctoral students.
Expand
Zhi Tian
Zhi earned his PhD in Computer Science from the University of Adelaide, Australia, where he was mentored by Professor Chunhua Shen. In 2019, he was honored with the Google PhD Fellowship.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Zhi is currently a research scientist for Vision Generation Models at ByteDance Doubao, focusing on the development and comprehension of computer vision algorithms. He has authored multiple highly-cited papers presented at leading conferences and in prestigious journals, amassing more than 11,000 citations on Google Scholar.
Expand
Zhuo Chen
Zhuo obtained his PhD from Columbia University in 2017 and worked as a core application scientist at Microsoft after graduation.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
He currently leads the Audio Generation Research team at ByteDance Doubao. With more than 100 research papers and patents, he pushed the frontier of multiple audio technologies and has made significant contributions in various areas including voice generation, recognition, and translation, voice separation and enhancement, speaker recognition and diarization, self-supervised learning for voice, multi-channel processing, as well as in the development of open-source voice data sets.
Expand
Chenggang Li
Chenggang obtained his bachelor's and master's degrees from the Department of Mechanical Engineering at Tsinghua University.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
He served consecutively as the head of ByteDance's Toutiao Webpage Search team and TikTok Video Search technology team. He developed ByteDance's search system from scratch, making significant innovations and breakthroughs in ranking architecture and algorithms, multi-language, and multimodal relevancy. Thanks to his contributions, ByteDance was able to deliver the leading search experience for Chinese-based search. Currently, he leads the Large Language Model Pre-training team at ByteDance Doubao, focusing on areas including data cleaning/synthesis/blending, associative learning and curriculum learning, training algorithms, and scaling capability.
Expand
Wanjun Zhong
Wanjun graduated in 2023 from the "Sun Yat-sen University - Microsoft Research Asia Joint PhD Program" and obtained a PhD in Computer Science, under the supervision of Dr. Ming Zhou, Professor Jian Yin, and Professor Jiahai Wang. During her PhD, Wanjun was awarded the 2021 MSRA Fellowship, the National Scholarship for PhD Students, and the CVPR Challenge Championship.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Wanjun is currently a researcher in the Doubao Seed LLM team at ByteDance, focusing on super alignment and inference of large language models. AGIEval, an LLM evaluation set she developed, is used for model evaluations by major mainstream companies.
Expand
*04
Personal Growth and Rewards
Access to ample computing power, vast datasets, and rich application scenarios
/ Ample computing power and datasets coupled with nimble team collaboration, enabling you to concentrate on technical research
/ Rich application scenarios that enable hundreds of millions of users to experience your findings
/ Abundant opportunities to explore key research areas, ensuring long-term goals are not compromised by short-term metrics
/ Large investment in AI foundation models and agile decision-making, helping ideas quickly come to fruition
Fully empower young individuals with a strong focus on professional growth
Highly competitive packages complemented by outstanding rewards
Access to ample computing power, vast datasets, and rich application scenarios
/ Ample computing power and datasets coupled with nimble team collaboration, enabling you to concentrate on technical research
/ Rich application scenarios that enable hundreds of millions of users to experience your findings
/ Abundant opportunities to explore key research areas, ensuring long-term goals are not compromised by short-term metrics
/ Large investment in AI foundation models and agile decision-making, helping ideas quickly come to fruition
Fully empower young individuals with a strong focus on professional growth
/ Experience full empowerment and trust that allow you to steer key research projects
/ Receive support from top mentors, who can help you identify the most critical and high-value research topics in the field
/ Get interviewed by mentors directly and participate in the long-term training program that lasts over a year, granting you the freedom to delve deeply into the research topic
/ Explore opportunities to communicate both internally and externally, exchanging insights on industry trends with leading experts
Highly competitive packages complemented by outstanding rewards
/ Top pay for top talents
/ Excellence is recognized and rewarded — your exceptional performance will never go unnoticed
/ Join us at any time and grow together with a company that's invested in your future
*05
Top Seed - Recruitment Process
The Top Seed program is open for both fresh graduates and research interns all year round. Students can apply according to their individual circumstances.
Application/Referral
Interview
Offer
*06
Top Seed - Q&A
Graduate Recruitment
Internship Recruitment
QWhat is the difference between the Top Seed Talent Program and the ByteDance Soaring Star Talent Program?
A
Both talent programs are aimed at PhD candidates graduating in 2025, with different research directions. If you aspire to work in fields such as machine learning, artificial intelligence, foundation models, computer vision, as well as audio and video generation, choose the Top Seed Talent Program. If you are keen on diving into fields such as AI applications, search, recommendations, advertising, AI for Science, AI Safety, robotics, privacy and security, hardware, video architecture, and engineering architecture, go for the ByteDance Soaring Star Program.
QWhat is the application mechanism like for the two talent programs?
A
In the Top Seed Talent Program, an applicant is granted one chance to apply, while in the ByteDance Soaring Star Program, an applicant has two opportunities to apply. The application opportunities for the two talent programs are independent, and we will prioritize the position that is applied for first. Moreover, applications for both talent programs will not impact the regular application process for the 2025 campus recruitment. We welcome exceptional individuals to join us!
QIf I am a candidate for the class of 2025 and have already received an internship offer from another team, can I still apply for a position under the Top Seed Talent Program?
A
A candidate can only be in the process for one position at a time. In special circumstances, please contact the HR of the current position for assistance.
If the answers above do not resolve your question, or if you would like to get more information, feel free to contact us via email attopseed@bytedance.com.