Production Engineer, Associate
Skills:Python,Bash,FIX Protocol,Linux,Observability
I deploy and operate infrastructure across cloud regions and data centers for teams where downtime is expensive: trading desks, robotics fleets, biotech pipelines. The goal is always the same. Provision cleanly, fail predictably, stay understandable.
Skills:Python,Bash,FIX Protocol,Linux,Observability
Skills:AWS EKS,Kubernetes,Helm,Terraform
Skills:AWS Lambda,DynamoDB,EventBridge,Python,Cloud Build
Skills:CloudFormation,Go,EC2,EventBridge,SNS
Skills:GCP,Pub/Sub,Dataflow,GKE Anthos,Kubernetes
A single construction job carries thousands of pages of contracts, specs, drawings, RFIs, and submittals. The answer to any given question is almost always in there. Finding it means reading, and the people who need the answer have the least time to read.
I built the product end to end: the retrieval and reasoning layer, the API, the interface teams actually work in, and the infrastructure it runs on.
Documents land in S3 and are indexed twice: dense embeddings for meaning, BM25 for the exact clause numbers and part codes that embeddings lose. Questions are answered against the source rather than a summary, so every answer traces back to the page it came from. Teams raise RFIs, log action items, and generate daily reports without leaving the documents behind.
Running in production on AWS: a FastAPI and PostgreSQL backend behind a Next.js frontend, transactional mail through SES, Cloudflare in front, and Sentry watching both halves. It ships continuously and serves real project teams.
Skills:FastAPI,Next.js,PostgreSQL,AWS,Hybrid RAG,TypeScript