Large Model Application Algorithm Research Scientist-International Content Security Algorithm Research-Soaring Star Talent Program

ByteDance
Singapore
Workplace: OnsiteFull timeFunction: Data Science & Machine LearningEducation: phdSkills: ["Problem-solving","Analytical skills","Communication","Collaboration"]

Develop and iterate machine learning models and information systems for international content safety, including foundational large models used across content moderation, search, and recommendation. Lead research on improving LLM reasoning via reinforcement learning, reward model design, stable RL training without expensive SFT data, expanding reasoning from math/code to natural language, and boosting reasoning efficiency through approaches like knowledge distillation and Long-CoT techniques.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
ByteDance
ByteDance
1 month ago

Large Model Application Algorithm Research Scientist-International Content Security Algorithm Research-Soaring Star Talent Program

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 30 days agoStatus: Live
Reposted: similar role first listed 1 month ago

Job Summary

Develop and iterate machine learning models and information systems for international content safety, including foundational large models used across content moderation, search, and recommendation. Lead research on improving LLM reasoning via reinforcement learning, reward model design, stable RL training without expensive SFT data, expanding reasoning from math/code to natural language, and boosting reasoning efficiency through approaches like knowledge distillation and Long-CoT techniques.
Location: Singapore
Workplace: Onsite
Employment Type: Full time
Job Function: Data Science & Machine Learning

Key Responsibilities

  • •Build and iterate on ML models and information systems for international content safety, including foundational large models for platform scenarios like moderation, search, and recommendation.
  • •Design reward models for RL-based reasoning that align with reasoning effectiveness and can adapt dynamically during training.
  • •Develop robust training strategies to stabilize RL training when high-quality SFT data is unavailable.
  • •Advance RL reasoning methods from math/code tasks to natural language tasks by innovating in data design and RL methodology.
  • •Improve reasoning efficiency while maintaining quality using approaches such as knowledge distillation and Long-CoT techniques for better cost-effectiveness.

Key Requirements

  • •PhD degree in Computer Science, Electronics, or other related fields.
  • •Extensive experience in ML/CV/NLP/recommendation systems, including participation in industry projects or competitions and/or related work.
  • •Publications in ML, data mining, AI, or large models (e.g., KDD, WWW, NIPS, ICML, CVPR, ACL, AAAI).
  • •Strong programming skills and hands-on contributions using Python/C++ (or other relevant languages).
  • •Strong problem-solving and analytical skills, with excellent communication and a collaborative mindset.
Experience:Machine learningNLPComputer visionRecommendation systemsLarge modelsReinforcement learning
Education:PhD / Doctorate
Skills:Problem-solvingAnalytical skillsCommunicationCollaboration
Tech Stack:Machine LearningMLCVComputer visionNLPRecommendation SystemsLarge modelsLLMsReinforcement learningRLProcess-based Reward ModelPRMMonte Carlo Tree SearchMCTSChain-of-ThoughtCoTLong Chain-of-ThoughtLong-CoTKnowledge distillationPython

Company Brief

ByteDance
Develops consumer internet and content platforms, including TikTok and other apps for short-form video, news, and entertainment. It also builds advertising, commerce, and creator tools that connect audiences, brands, and publishers across global markets.
Industry: Digital Media
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Established Company
Headquarters: Beijing, China
Founded: 2012
WebsiteLinkedIn