About the company
Our client is an artificial intelligence company.
The role
Raydar is recruiting for this role on behalf of our client. Lead applied machine learning research projects from initial idea through model training, testing and production rollout. Build the tooling that supports experimentation and contribute to the technical direction of a young research group.
What you'll do
- Fine-tune large language and multimodal models after pretraining, using reinforcement learning and related techniques.
- Investigate how autonomous software agents can stay effective on extended tasks, including planning, retaining context, recovering from errors and deferring to people at the right moments.
- Create simulated software environments, data generation pipelines and test suites used to train and measure agents.
- Develop approaches that let agents interpret and keep track of what users are trying to achieve across lengthy sessions.
- Plan careful experiments and choose the problems most worth pursuing.
- Move promising findings into production features.
- Contribute to setting research priorities and engineering direction for the group.
Requirements
What we're looking for
- Outstanding skill in both machine learning research and engineering.
- Deep expertise in at least one area such as model fine-tuning, reinforcement learning, agents, reasoning, memory and context handling, computer use, evaluation or human-agent interaction.
- Sound judgment in designing experiments and a habit of delivering complete working systems instead of one-off demos.
- Ability to shift smoothly between exploratory research, large training runs and production engineering.
- A track record of owning significant research work, for example a notable model, agent system, dataset, publication or open-source contribution.
- Self-driven, ambitious and at ease working on open-ended problems with no known answer.
- Roughly 3 to 6 years of research or engineering experience, with time at a leading AI lab valued.
Bonus points
- Direct experience with reinforcement learning for model fine-tuning, agents or long-running systems.
- A solid record of publications, open-source work or released benchmarks.
- Experience building training environments, data pipelines or test harnesses.
- Exceptional competitive achievement, such as in olympiads or quantitative fields.
Benefits
Compensation and benefits
- Base salary: USD 250,000 to 500,000 per year
- Equity
Location and work model
- San Francisco, CA, United States
- On-site, in office
- Full-time