AI Architect/ Engineer

Upwork Job Post Title: AI Engineer / Architect ; MCP, Agents, API Integrations & Production AI Infrastructure We are looking for a hands-on AI Engineer / AI Systems Architect to help build production-grade AI agent infrastructure. This is not a simple chatbot or demo project. We need someone who can design and build reliable, secure, cloud-native AI systems that connect agents to real business tools, APIs, workflows, and data sources. Required Skills You must have strong experience with: * Model Context Protocol * Building MCP servers * Tool schemas and agent integrations * AI agent workflows and tool use * Node.js/JavaScript and/or Python * REST APIs and OAuth2 * Token refresh, expiry handling, and rate-limit handling * Structured data reads/writes * Retries, queues, idempotency, and deduplication * Cloud deployment: AWS Lambda, GCP Functions, Cloud Run, or containers * Secrets management and secure credential storage * Git, CI/CD, automated testing * Logging, observability, audit trails, and alerting What You Will Build You will help design and implement systems where AI agents can safely interact with business tools and APIs. The system must be: * Secure * Reliable * Deterministic * Testable * Auditable * Cloud-native * Not dependent on one local machine We care about production quality: evals, regression testing, failure recovery, structured outputs, clear error handling, and repeatable workflows. Bonus Experience Highly valued: * Anthropic Claude API * Claude tool use / agent workflows * Prompt evaluation and skill design * Cost and latency optimisation * Google Apps Script * Google Workspace automation * Gmail, Calendar, Drive, Sheets APIs * Multi-agent orchestration * AI governance, auditability, and control frameworks Ideal Candidate You should be able to explain how you would: * Build an MCP server from scratch * Connect an AI agent safely to third-party APIs * Handle OAuth2 and token refresh * Prevent duplicate writes/actions * Test and evaluate AI workflows * Make agent outputs more deterministic * Log and audit every action * Deploy safely to cloud infrastructure To Apply Please include: 1. Relevant MCP or AI agent work you have built 2. Your preferred stack 3. API integrations you have delivered 4. How you handle testing, retries, failures, and reliability 5. Experience with Claude API or Google Workspace automation, if any.