Founding Enginee

• Own the core backend modules. Take over live production systems, read into them fast, refactor and harden them, and keep them reliable in production.
• Lead new backend development. Own new features from 0→1 and 1→100: API and service design, data models, business logic, background jobs, caching, and reliability. Requirement to shipped, tested, deployed, maintained.
• Build the AI layer. Ship LLM and agentic features (chat, agents, RAG, retrieval) and make non-deterministic output reliable, with real attention to prompting, evals, context windows, cost, and latency.
• Grow into the backend lead. Set technical direction for the core backend and new features, and mentor the team over time.

What We're Looking For

Technical Must-Haves

• Deep core backend, run in production. API and service design, databases, data models, business logic, background jobs, caching, and reliability — with real fundamentals underneath (system design, data structures, concurrency, DB internals), not just framework familiarity. Built and operated for real users, not just prototypes. Strong in the stack the inherited modules run on (Python / Go / Node, SQL + NoSQL).
• Event-driven and async systems; messy integrations. Has built and operated event-driven, async backends in production (message queues, async pipelines, scheduled jobs and crons) and is comfortable with messy third-party integrations: email (IMAP / MS Graph), WhatsApp and messaging APIs, ERPs.
• Modular design; inherits and hardens existing code. Builds loosely-coupled, composable modules with clean interfaces, reads into someone else's code fast, and refactors and hardens rather than rewrites from scratch.
• End-to-end, T-shaped feature ownership. Owns a feature (Core Backend to Broad Module Development). Has taken features from requirement to maintained solo, both 0→1 under ambiguity and 1→100.
• AI-native working (force multiplier). Uses AI tooling to multiply engineering throughput.
• AI and LLM feature-building (hard requirement). Has shipped AI and LLM features into a product (chat, agents, RAG, retrieval) with a real grasp of prompting, evals, context windows, and cost and latency tradeoffs. A hard requirement, not a nice-to-have.