Backend Engineer — LLM Infrastructure

You will work on the always-in-path FastAPI proxy that sits between every customer request and every provider. That means guardrails, prompt management, A/B testing, cost attribution, rate limits, and streaming. You will own the parts of the request lifecycle that have to stay correct under real production load.

• Extend the Nemo Backend FastAPI proxy with new guardrails and features
• Own streaming, retries, and provider failover correctness
• Build and maintain cost attribution from x-nemo-request-cost
• Profile and tune hot paths — every millisecond is in the user-facing latency budget
• Harden multi-tenancy isolation at the request layer

• 5+ years of backend Python in production
• Deep experience with asyncio and high-concurrency services
• Comfortable with Postgres, connection pooling, and query optimization
• Production experience with streaming APIs or proxies

• Prior work on LLM APIs, model gateways, or SSE streaming
• Experience with LLM routing engines or model gateways
• Performance profiling and flame graph literacy