Composio
Forward Deployed Research Engineer
About this role
Composio seeks a Forward Deployed Research Engineer to work directly with enterprise customers on the cutting edge of AI agent infrastructure. You'll transform complex APIs into agent-accessible tools, ship production code, and help define best practices for an emerging field—embedding with customers and turning research questions into deployed solutions.
What you'll do
- Own enterprise accounts end-to-end from discovery through implementation and release
- Ship production code for custom tools, schemas, and per-customer patches in the toolkit monorepo
- Convert large enterprise APIs into agent-usable tool surfaces and build evaluation harnesses
- Conduct weekly customer syncs, manage timelines, and handle escalations
- Design tool granularity, scope, and descriptions to optimize model navigation
- Build repeatable QA processes and enable self-service tool patching
What they're looking for
- Production integration engineering and enterprise API expertise
- OAuth 2.0 and authentication/scope design experience
- Python or TypeScript proficiency
- Large language model application development
- Technical account ownership and customer communication
- API platform or iPaaS background
- Systems design under pressure
- Clear technical writing and explanation
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Composio
Composio builds agent-to-tools infrastructure that enables AI agents to interact with external systems and APIs. The company is hiring AI engineers, deployment specialists, growth engineers, and customer-focused research engineers to expand its platform, improve production deployments, drive adoption, and work directly with enterprise customers.
- Website
- composio.dev
Likely interview questions
- Walk us through a time you took an enterprise API from discovery to production integration. How did you decide what to expose to users, and what trade-offs did you make?
- Tell us about a complex OAuth or auth flow you've implemented. How did you handle edge cases around scopes, consent models, or multi-workspace tokens?