Research, Finetuning Science
San Francisco
Workplace: HybridFull timeUSD 350,000 - 475,000 annuallyFunction: Research & Scientific (R&D)Education: bachelorsSkills: ["Communication","Debugging","Distributed training","Probability/stats fundamentals","Writing"]Advance fine-tuning and frontier post-training techniques to improve Tinker's training defaults and the underlying post-training engine. You’ll work on LoRA and parameter-efficient fine-tuning, increase stability and reliability of large-scale fine-tuning and RL runs, and ship research that becomes product primitives and cookbook recipes. Collaborate with internal research teams and contribute to open-source resources through papers and technical blog posts.
Loading
Loading job details...
Preparing the role view and application actions.

