SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Composio is building infrastructure that enables AI agents to interact with workplace tools like GitHub, Gmail, Notion, and Salesforce. The company raised a $25M Series A from Lightspeed with backing from founders at Vercel, HubSpot, and others, serving customers including Brex, Glean, and Zoom.
This is an applied AI engineering role focused on building agentic systems that automate customer support workflows. You will work directly with customers to understand their pain points, then design and ship AI-powered solutions that handle support functions end-to-end: intake, classification, context gathering, diagnosis, response drafting, remediation, and verification.
Key responsibilities include:
- Building agentic workflows that automate support functions with proper guardrails, permissions, and escalation paths
- Creating tool-using agents that safely inspect system state and execute bounded recovery actions
- Developing comprehensive eval suites for diagnostic correctness, resolution quality, and safe escalation
- Implementing human-in-the-loop systems that make ownership, uncertainty, and failure states explicit
- Designing observability and feedback loops so agents improve from real support cases
- Identifying product improvements that eliminate entire classes of customer issues
You'll start with customer outcomes and work backward to determine whether the solution requires model tuning, workflow changes, tooling, or product changes. You'll own hard issues end-to-end when direct engineering is needed, reproduce and isolate problems, ship fixes, and verify resolution with customers. You'll partner with Support, Field Development Engineers, Product, and Engineering teams to identify work that should become software.
Required: 2–4 years of professional software engineering experience with production systems using language models, tool calling, retrieval, structured outputs, multi-step workflows, or agent memory. Strong backend, platform, or integration engineering background with ability to debug across SDKs, HTTP, OAuth, webhooks, queues, data stores, and third-party APIs. Customer-adjacent mindset with clear communication skills and ability to turn vague problems into systems. Workflow and product judgment to model support functions as states, tools, permissions, and escalation rules.
Optional: TypeScript or Python production experience, AI agents/copilots/workflow automation background, developer platform or SRE experience, OAuth/webhooks/Postgres familiarity, support system experience (Plain, Zendesk, Intercom, Salesforce), or public technical writing.
