Computer vision / Cybord AI
A ten-minute pipeline. A one-minute result.
Taking AI-powered circuit-board inspection from a proof of concept to production.
The challenge
Circuit-board inspection brings several difficult tasks into one workflow: scanning boards, analyzing components, extracting markings, and comparing multiple boards. The engineering challenge was to connect those capabilities into a production-ready backend while bringing processing time down.
The engineering
- Engineered a microservice backend for board scanning, component analysis, and multi-board comparison.
- Used task-graph parallelization to accelerate a pipeline that previously took roughly ten minutes.
- Connected computer vision, OCR, and image processing with a Python stack using Kafka, gRPC, FastAPI, Redis, and OpenCV.
- Developed polarity, rotation, body-size, and missing-component comparison logic, plus a CAD-alignment service built on FastAPI, Celery, and Kafka.
- Packed component images into one bundle per board and served each image through Nginx HTTP range requests, with byte offsets stored in SingleStore.
- Added real-time WebSocket alerts with a Redis pub/sub relay across replicas, and OpenTelemetry tracing with workflow IDs carried in Kafka headers.
The result
Pipeline processing fell from approximately ten minutes to one minute. The proof-of-concept backend was accepted by SpaceX and is now in production.
05.Open a connection
Production ML, computer vision, AI infrastructure, or a team looking for an engineer who ships. Tell me what you have in mind. Three fields, one conversation.
Fill out the form → hit send → I’ll reply to your email.
↓ download résumé · PDF