•Improved and built product offerings, helping scale the company to ~6M ARR until eventual acquisition by Salesforce
•Architected and owned core backend services handling up to 1.1k peak RPS, ensuring low latency and high availability across search and recommendations platforms.
•Built and productionized a rules-driven recommendation engine supporting catalog segmentation, boosting/burying, filtering, and multiple recommendation groups; load-tested and deployed to high-traffic enterprise retailers.
•Designed and implemented a microservice architecture leveraging LLMs for intelligent search query evaluation and rewrites, enabling self-improving search behavior and stabilizing production traffic across 10+ customers.
•Reduced search and recommendation latency by 80–90% through multi-layer caching (API, inference, session, catalog), eliminating thundering-herd behavior during peak and holiday traffic.
•Led performance investigations and incident response across distributed systems (Elasticsearch, TorchServe, Pub/Sub, GCP), authoring postmortems and implementing durable reliability improvements.
•Acted as technical owner for multiple customer launches and demos, coordinating backend APIs changes in Golang, UI updates in React, infrastructure configs for deployments to Kubernetes, and model deployments while unblocking teammates and driving features from design to production.
