AI product Engineer | Agentic AI, SaaS & API Integrations

We’re looking for an experienced developer to build and improve a SaaS platform that connects AI agents, business data, and third-party tools into reliable workflows. The role involves translating business requirements into working features across the frontend, backend, and AI layers. You should be comfortable making architectural decisions, investigating technical problems, and taking features through testing and deployment. Scope: - Develop AI agents with tool calling, conversation memory, structured outputs, and multi-step workflows. - Connect business tools through APIs, webhooks, and secure authentication. - Build RAG and document-search features where needed. - Develop backend services and interfaces for reviewing agent activity and managing workflows. - Support multiple customer accounts with isolated data and permissions. - Implement human approvals, duplicate prevention, retries, and recovery from failed operations. - Improve response quality, execution speed, memory usage, and API costs. - Document the architecture, integrations, and deployment process. Relevant experience: - Python/FastAPI or Node.js/TypeScript - React or Next.js - OpenAI, Claude, or Gemini APIs - Agent orchestration using LangGraph or similar tools - Databases, vector search, and structured data processing - n8n or custom workflow automation - SaaS development, authentication, and production deployment Deliverables: - Working features connected to the agreed tools and data sources - Tested workflows covering successful execution, invalid inputs, and API failures - Source code, configuration instructions, and technical documentation - A demonstration and handover for each completed milestone The first milestone will focus on one agreed workflow from input through agent actions to the final result. Additional features will follow after testing and acceptance. When applying, please share: 1. Two relevant AI agent or SaaS projects and your personal contribution. 2. An example of handling tool failures, duplicate actions, or conversation memory. 3. How you decide between deterministic automation and an AI agent. 4. Your availability, hourly rate, and preferred engagement structure.