Composio
Member of Technical Staff - Research Engineer
sf · On-site · Staff · Full-time · Posted Feb 10, 2026
About the job
Develop agent infrastructure by evaluating and improving models’ long-horizon tool execution, tool and plan search, and agentic systems.
What you'll do
- Build evaluations using real tool-calling data to measure model performance in long-horizon tool execution.
- Develop search for semantically similar tools and cached execution paths and plans.
- Train agentic harnesses and improve them through supervised fine-tuning and reinforcement learning on agentic traces and app sandboxes.
What they ask for
- Independently execute research goals and prototype and test experiments quickly.
- Work with product and engineering teams to move research ideas into production quickly.
- Write documentation well and explain complex ideas clearly.
- Build trust and acknowledge what you do not know.
AI focus
Research and improve AI agents’ tool execution through evaluations, semantic search, agentic harness training, supervised fine-tuning, and reinforcement learning.
Model-written summary · the original posting is authoritative