Michael Lipski

Early engineer at a venture-backed AI SaaS startup focused on platform scalability, enterprise integrations, and LLM evaluation tooling. Architected high-throughput microservices on GKE supporting 10M+ customer records. Built an end-to-end automated LLM evaluation and QA benchmarking system across Go and React, including F1 scoring, rate-limited throttling, and an interactive review dashboard to accelerate audit turnaround by ~80%. Re-architected batch ingestion from 10k to 1M+ records per batch using dynamic row/size chunking, a Go worker mesh, streaming row-write flushing, gRPC/Protobuf, and PostgreSQL to prevent OOM failures. Created an event-driven integration framework syncing near-real-time data from Zendesk and Medallia using React OAuth/credential flows, a Go webhook receiver, RabbitMQ queues, and resilient consumer retry policies. Scaled distributed taxonomy curation with Kubernetes workers and admin tooling for clustering/deduplication across multi-tenant dashboards. Implemen

Full stack engineer for a Series D AI-driven B2B procurement platform. Developed enterprise workflows with React, GraphQL, and Python microservices deployed on AWS ECS/EKS. Served as Technical Lead to deliver a key customer-facing full stack feature enabling provider competition, coordinating with product, design, and QA. Improved production stability by managing CI/CD lifecycles through Harness and participating in post-mortems to implement preventative logic. Architected complex PostgreSQL schemas to capture user-interaction telemetry and transform raw event data into datasets for product modeling.