Remote Research Scientist - LLM Foundation Models & Reasoning

Blockchain Academics; Hong Kong; Remote; HKD 1,200,000 - 1,800,000; Full time

Benefits offered by this job

Remote-first work environment

Competitive salary

Career growth and learning

Job summary

Binance is seeking a highly skilled Research Scientist/Engineer to advance reasoning and planning in large foundation models. You will drive development across data acquisition, supervised fine-tuning, reward modeling, and RL, while designing robust evaluation methods.

You will enable models to interact with tools, APIs and code interpreters, and build agents and multi-agent systems for complex real-world challenges.

Qualifications

• Proficiency in RL, LLM and familiarity with large-scale model training.
• Strong data structures and fundamental algorithm skills; fluent in Python or C++/Java.
• Experience with influential projects or papers in RL, NLP or Deep Learning.

Responsibilities

• Reasoning and planning for foundation models across data acquisition, model evaluation, SFT, reward modeling, and RL.
• Synthesize large-scale, high-quality data via rewriting, augmentation and generation for pretraining, SFT, RL stages.
• Solve complex tasks using system 2 thinking and advanced decoding strategies like MCTS and A*.
• Develop robust evaluation methodologies to assess model performance at various stages.
• Teach models to use tools, interact with APIs and code interpreters; build agents and multi-agent systems.

Skills

RL & LLM

Large-scale model training

Python / C++ / Java

Data structures & algorithms

Research publications (RL/NLP/DL)

Job description

Binance is seeking a highly skilled Research Scientist/Engineer to advance reasoning and planning in large foundation models. You will drive development across data acquisition, supervised fine-tuning, reward modeling, and RL, while designing robust evaluation methods.

You will enable models to interact with tools, APIs and code interpreters, and build agents and multi-agent systems for complex real-world challenges.