Backend Engineer II Sequoia
Built backend services for the Merit Planning platform in Sequoia OS, supporting compensation and workforce planning for more than 200 enterprise customers with strict multi-tenant isolation and AuthN/AuthZ. Designed and shipped six FastAPI microservices and model-serving APIs using MongoDB and AWS for employee-data and compensation-review workflows, handling more than 1 million records with p95 latency below 150 ms.
Integrated LLM-powered compensation analytics and automated workspace insights using AWS Bedrock and LangChain. Created an offline and online LLM evaluation harness with golden datasets, LLM-as-a-judge scoring, and CI/CD regression gates for model, prompt, and quantization updates.
Instrumented backend and AI inference services with OpenTelemetry, Prometheus, and distributed tracing to monitor tenant token spend and latency drift, reducing mean time to detect incidents by 40%. Partnered with product, frontend, and platform teams to deliver enterprise features while maintaining 99.9% service availability.