Founding Staff Engineer

FOUNDING STAFF SOFTWARE ENGINEER — AGENTIC SYSTEMS & BACKEND Leela is building continuous, agentic regulatory compliance for financial services — replacing manual obligation tracking with AI that reasons about regulations and produces verifiable, trustworthy results. ABOUT THE ROLE As a founding engineer, you'll build the core of our platform: multi-agent AI systems whose outputs are grounded in a deterministic engine, plus the backend that runs them in production. This is deep work — agent design, formal/rule-based reasoning, and reliable systems — for someone with strong CS fundamentals who sets the technical bar. You must have shipped production systems for 7+ years for this role. US Citizen / Permanent Residents only WHAT YOU'LL DO • Design and build multi-agent AI systems — and the evaluation harnesses that keep them accurate and honest. • Design and build a custom domain-specific language (DSL) for the regulatory domain — encoding rules into a deterministic engine with verifiable outputs. • Build reliable, well-typed backend services and data models that orchestrate AI workloads in production. • Work across LLM providers (Anthropic, OpenAI, Google), including cost, evaluation, and model-selection strategy. • Set engineering standards — strong typing, fail-loud error handling, and a codebase the next engineer can read. WHAT WE'RE LOOKING FOR Must-have • Strong CS fundamentals: data structures, concurrency, state machines, and distributed-systems reasoning. • Production experience in TypeScript/Node (or a comparable typed language, ready to go deep in TS). • ML / AI experience — hands-on with AI/LLM SDKs and agentic frameworks (Vercel AI SDK, Anthropic/OpenAI SDKs, LangGraph or similar); research-ML background optional. • Fluency with relational databases / PostgreSQL. Nice to have • Compilers, DSLs, formal methods, or other rule/logic-based systems. • Background in compliance, regtech, fintech, or another high-stakes correctness domain. SUCCESS LOOKS LIKE Within the first year: owning a core subsystem end-to-end and setting the architecture and standards for how we build agentic systems.