<p>We're looking for a Senior Integration Engineer to build high-performance, secure, and production-grade API layers, custom Model Context Protocol (MCP) servers, and real-time data delivery services optimized for both human applications and autonomous AI agents. This role requires experience operating at high throughput — thousands of requests per second — and building integration layers that AI agents rely on directly.</p><h3>Requirements<p></p><p><strong>Must-Have</strong></p></h3><ul><li><strong>Overall Experience: 8+ years in Backend Software Engineering, Distributed Systems Architecture, or Enterprise Platform Engineering</strong></li><li><strong>Resiliency Engineering: 4+ years architecting mission-critical systems requiring "four-nines" (99.99%) availability, rigorous traffic shaping, and self-healing systems</strong></li><li><strong>AI &amp; Agentic Systems: 2+ years building production-grade integration layers, semantic context gateways, or RAG endpoints consumed by LLMs and autonomous AI agents</strong></li><li><strong>Proficiency in Advanced API Design &amp; Context Optimization</strong></li><li><strong>Production-level proficiency in one or more of: C# (.NET Core), Java, Python, or Node.js/TypeScript</strong></li><li><strong>Proficiency in Enterprise Scalability &amp; Traffic Management</strong></li><li><strong>Proficiency in Lock &amp; Contention Mitigation (API &amp; Database Tier)</strong></li><li><strong>Proficiency in Hardened Security, Governance &amp; AI Isolation</strong></li></ul><h3><strong>Preferred Experience</strong></h3><ul><li><strong>GraphQL and dynamic JSON/gRPC filtering layers (zero-waste data fetching)</strong></li><li><strong>Custom MCP server development for LLM context exposure</strong></li><li><strong>Heterogeneous schema merging for LLM tool-calling (e.g., relational DBs, analytics platforms like Pendo/Hotjar, observability tools)</strong></li><li><strong>Semantic token-aware API pagination and streaming</strong></li><li><strong>Context-aware gateways for conditional object graph hydration based on agent intent</strong></li><li><strong>Edge performance: CloudFront Functions, Lambda@Edge, HTTP/3, WebSockets</strong></li><li><strong>PostgreSQL and OpenSearch/Elasticsearch access layers with sub-millisecond caching</strong></li><li><strong>Reactive streams and rate-limiting for Amazon MSK (Kafka), SQS, and HTTP endpoints</strong></li><li><strong>Circuit Breakers, Bulkheads, and Retry-with-Exponential-Backoff (Polly, Resilience4j, or equivalent)</strong></li><li><strong>Amazon MemoryDB / Redis OSS / Valkey for rate limiters, session caches, and semantic query caching</strong></li><li><strong>Optimistic Concurrency Control (OCC) via eTags/versions; Pessimistic Locking only where strictly necessary</strong></li><li><strong>CQRS separating high-volume writes from intensive read queries in PostgreSQL</strong></li><li><strong>Asynchronous write delegation for high-contention API writes to Kafka/MSK</strong></li><li><