Sergei Smekhunov

Delivered and operated high-availability delivery systems on Cloudflare Workers (serverless, Lambda-equivalent, TypeScript, hono) handling 1M+ users daily with ~43ms average latency. Improved observability standards (OTLP, TypeScript, Grafana) to ensure consistent tracing across distributed serverless services. Built configuration management services (BullMQ, NestJS, Prisma, PostgreSQL) as RESTful microservices to keep company user configuration synchronized across multiple systems. Designed reinforcement learning solutions using MAB-like algorithms to help users automate experiment adjustment and distribution, reducing the mental overhead of supporting marketing campaigns. Led tasks with small cross-functional teams (2–8 people), refined testing and knowledge-sharing practices (Jest standards), and improved overall coverage and velocity. Implemented an AI-based recommendations system (Langfuse, Temporal, TypeScript) to suggest personalized, multi-level configuration templates based

Led a cross-functional team of 8 to create a fleet of microservices collecting, refining, and providing machine/asset data for downstream teams and end users. Built and scaled Node.js microservices using NestJS and Express to handle global machine operations with 99.99% uptime. Built and improved an image processing microservice (Python) to handle and prettify machine data. Established coding and development standards adopted across three teams and mentored junior and senior engineers. Built high-capacity data search and clustering solutions using Elasticsearch to improve availability and usability of machine data.

Improved a medical data processing pipeline (Python, AWS, Kubernetes, Argo) handling 300GB+ data chunks, reducing processing time by 30%. Built an RBAC (role-based access control) system (TypeScript, NestJS, PostgreSQL) to regulate data access for users from different hospital systems. Maintained and optimized a statistical data-serving backend (TypeScript, Python, R, gRPC, Elasticsearch, MongoDB, Docker), streamlining gRPC chains and narrowing sources of truth. Improved a UI for displaying statistical data (Angular, Plotly) to make it more stable and to add reliable user and permission handling.