Solutions Architect

This is not a demo role. As Solutions Architect, you are the technical owner of deals with developer-led companies and enterprises investing in agents. You walk into a customer, learn how their identity, data, apps and governance fit together, and design an architecture that survives their security review. You build the proof of concept yourself, and you hand a working deployment to our forward deployed engineers after close. Customers should come to see you as an adviser on how to move their agents from proof of concept to production, not as a vendor.

What you will do

• Own the technical win. Partner with our account executives to lead discovery and evaluation with heads of AI, CTOs and architects.
• Design the architecture. Decide how a customer’s agents connect to their apps: SDK, MCP or MCP Gateway; managed or custom OAuth; cloud, VPC or self-hosted.
• Build the proof. Build proofs of concept end to end in TypeScript or Python, against the customer’s real apps.
• Clear security review. Work through token custody, data flow, SSO/SCIM and audit logs with platform, identity and security teams.
• Build for the team. Create demos, reference architectures and sample agents that the wider sales team can reuse.
• Close the loop with the product. Bring repeated customer problems to product and engineering when they should become platform features. If a small fix will unblock a deal, ship it yourself., * Agent builder. You have built agents with the Claude Agent SDK, OpenAI Agents SDK, LangGraph, Mastra or similar, and you have felt the pain of giving them access to real apps.
• Backend engineer. You have shipped and run production services, and you know what it takes to observe and scale them.
• Identity and security engineer. You know OAuth 2.0, OIDC, SAML, SCIM and token lifecycles in depth.
• Integration engineer. You have connected enterprise systems with iPaaS, ETL or API gateways, and you see tool calling as the next version of that work.
• Founder. You have built a developer product and sold it to technical buyers yourself.

Customer Relationship Management (CRM), MCP - Microsoft Certified Professional, Authentication, WABI (Windows Application Binary Interface), Investment Capital, Juniper Networks M-Series, Microsoft FAST, Product Demonstration, Proof of Concept, Architectural Design, * 4+ years in solutions engineering, solutions architecture, forward deployed engineering or software engineering, ideally at a developer tools, infrastructure or AI company.

• Applied engineering. You ship production TypeScript or Python
• AI-native. You build with Claude Code, Codex or Cursor every day, you have shipped an agent that calls real tools on real accounts.
• API and authentication depth. You understand auth in detail - OAuth 2.0, SAML, SCIM, SSO.
• Enterprise experience. You have taken a technical evaluation through security review, and you work as well with a staff engineer as with a CIO.
• Clear communication. You write precise architecture documents and prompt follow-ups, and you explain complex products simply, to engineers and executives alike.
• Ownership and urgency. You are comfortable where the playbook is still being written, and you create structure without waiting for it.
• Trustworthiness. You build trust with customers, ask direct questions, admit what you do not know, and follow through.

San Francisco, in office · Full-time

An agent is only useful when it can act. It has to read the inbox, update the CRM, open the pull request and file the ticket, on the right user’s account, with the right permissions. Most agents still cannot do this reliably in production.

Composio is the action layer that makes it work. We connect agents to 1,500+ apps, with managed authentication, tool definitions models can use, and the controls an enterprise needs before it lets an agent near real data. Developers reach us through our SDK, MCP and the MCP Gateway.

Why Composio

• Traction. We serve more than 1B tool calls a month, and we have tripled our ARR since the start of this year.
• Customers. They range from early-stage YC companies to Wabi, Glean and Zoom.
• Backing. We raised a $25M Series A led by Lightspeed, with angel investors including Guillermo Rauch (CEO, Vercel), Dharmesh Shah (CTO, HubSpot) and Gokul Rajaram.
• Team. We are a small team of engineers working on hard problems, from context to search, and we ship fast., Solutions engineering is one path into this role, not the only one. If one of these describes you, we want to talk