SixSense

SixSense builds AI that inspects semiconductor wafers for defects, catching, classifying, and analysing failures that are invisible to the eye and extremely costly to miss. It is deep, niche, high stakes engineering: models in the loop, enormous image volumes, real fabs depending on the output.

I joined as a backend intern and it turned into the steepest growth of my career. I picked up real ownership early, learned to move through a large and mature codebase, and grew into someone the team trusts with load bearing systems. Fast, demanding, and the kind of place where the ceiling keeps moving up.

• Backend services on the core inference platform, designing and evolving APIs on a large Django and DRF codebase.
• Async processing with Celery: long running, queue driven pipelines.
• Vector search and embeddings with pgvector: on demand similarity instead of expensive re inference.
• Deployment and infra on GCP: Cloud Run, GCS, Cloud SQL.
• Hardware bound licensing: locking on prem deployments to a machine with signed, fingerprint derived licenses.
• An autonomous AI dev pipeline: a system that runs engineering tasks in isolation and hands the privileged writes back to a human.
• Performance and query work: chasing the slow paths and the subtle correctness bugs at scale.