Research Engineer, Agents
Core
Design and optimize agentic systems using LLMs to handle complex end-to-end tasks, including novel harness design, benchmarking, and model finetuning.
Role type
Senior Research Engineer (Agentic Systems)
Builds
Agentic systems, quantitative benchmarks, and evaluation frameworks for Claude
Domain
AI Safety, Large Language Models, Agentic Workflows
Deliverable
production ML models
Required skills
LLM-based agentic system development, software engineering, machine learning, prompt engineering, automated evaluation, data mix optimization
Preferred skills
Large-scale RL on language models, multi-agent systems
Technologies
LLMs, Claude, Python (implied by engineering context)
Responsibilities
Ideate and compare performance of different agent harnesses, design rigorous quantitative benchmarks for agentic tasks, assist with automated evaluation of models and prompts, work with product org to solve challenges applying agents, create and optimize data mixes for model training
Seniority
Senior, hands-on IC
