We're looking for engineers who can design and build production systems that combine strong software fundamentals with hands-on exposure to modern AI/ML tooling β specifically agentic AI workflows and Model Context Protocol (MCP) servers. You don't need to come from a pure AI/ML background; what matters is strong core engineering skills, sound architectural judgment, and the ability to pick up (or already know) how to integrate AI agents into real systems.
Strong grounding in modular design β clear separation of concerns, well-bounded components/services, low coupling / high cohesion
Ability to design systems that scale horizontally β statelessness where it matters, partitioning/sharding strategy, load distribution, avoiding single points of contention
Working fluency in core design patterns and principles (e.g., SOLID, domain-driven boundaries, event-driven/async patterns, idempotency, caching strategies, API versioning) and knowing when to apply vs. avoid them
Comfortable reasoning about tradeoffs: consistency vs. availability, latency vs. throughput, build vs. buy, synchronous vs. async workflows
For Lead specifically: expected to drive these principles β set direction in design reviews, push back on designs that won't scale, and mentor others toward this way of thinking rather than just applying it.
API & Real-Time Integration Best Practices Strong grasp of RESTful API design β resource modeling, versioning, pagination, idempotency for retries, proper status codes/error contracts
Practical experience with WebSocket-based real-time communication β connection lifecycle management, reconnection/backoff strategy, heartbeat/keep-alive handling, and graceful degradation when a socket drops
Understands tradeoffs between REST, Web Sockets, and event-driven/streaming integration patterns, and picks the right one for the use case
Follows security best practices for integrations β authentication/authorization (OAuth2, token scoping),input validation, rate limiting, and safe handling of external payloads
Designs integrations for observability and resilience β structured logging, timeouts, circuit breakers, and clear failure modes rather than silent drops