JOB-1942
A research lab dedicated to making language-model agents dependable and useful for individuals, developers and enterprises is expanding its core research team. Researchers here carry projects from first idea through experiments to working prototypes, alongside engineers and designers.
Scope
- Lead research on how LLMs reason, on post-training, and on agentic system design
- Invent methods that make autonomous agents more capable, reliable and safe in real settings
- Partner with platform and application engineers to turn research results into compelling products
- Plan and run experiments, benchmark models and study behavior to find failure modes and openings
- Follow progress in reasoning, multi-agent setups, RLHF, tool use and fine-tuning, and contribute papers or open-source code where suitable
- Help define the lab’s research culture and roadmap as an early member
About you
- Two or more years of LLM research on agents, post-training or evaluation, with publications
- A PhD or Master’s in computer science, machine learning, NLP or AI
- Original research you led, rather than applying existing methods to internal data
- Solid ML fundamentals in architectures, training recipes and experiment design, plus Python and PyTorch
- Papers at leading venues, where quality and citations matter more than the degree
On-site in the South Bay of Silicon Valley. Visa transfers considered, with sponsorship possible for strong candidates.
