Research Lead, Tinker, Fine-tuning Science
$350k - $475k • San Francisco • FullTime
Posted 1mo ago
Job Location
San Francisco
Tech Stack
Remote Work Policy
On-site
Employment Type
FullTime
Categories
AI Research Engineer
About the job
Thinking Machines is building AI that extends human will and judgment, with a focus on training frontier models and developing tools that allow users to customize AI for their unique needs. This role leads the Fine-tuning Science team, focusing on frontier customization techniques and the Tinker post-training engine. It's a player-coach position where you will remain hands-on with scientific research while also guiding the team, shaping its direction, and ensuring research findings are integrated into Tinker. You will collaborate with internal research teams, contribute to open science, and engage with external users.
Responsibilities
- Set the research agenda for advancing Tinker's quality, efficiency, and reliability.
- Lead, mentor, and develop a team of researchers, establishing high standards for experimental rigor.
- Conduct hands-on research in areas like LoRA, parameter-efficient fine-tuning, and their interaction with RL and post-training.
- Enhance the stability, efficiency, and reliability of large-scale fine-tuning and RL runs on Tinker.
- Represent the team's work externally through publications, technical blogs, and community contributions.
Requirements
- Bachelor's degree or equivalent experience in a quantitative field (CS, ML, Physics, Math) with strong theoretical and empirical grounding.
- Proven track record of leading research efforts and delivering results.
- Proficiency in Python and familiarity with deep learning frameworks (PyTorch, TensorFlow, JAX).
- Experience debugging distributed training and writing scalable code.
- Excellent communication skills for explaining complex technical concepts and building alignment.
- Strong interest in enabling custom AI models.
- PhD in a quantitative field or equivalent industry research experience (preferred).
- Experience with RLHF, RLAIF, preference modeling, or reward learning (preferred).
- Experience managing human data collection or large-scale annotation workflows (preferred).
- Contributions in alignment, data-centric AI, or human-AI collaboration (preferred).
- Experience with RL training stability techniques for large runs (preferred).
Benefits
- Generous health, dental, and vision benefits
- Unlimited PTO
- Paid parental leave
- Relocation support