Case studies.
Deployment work across three practice areas: industrial computer vision, enterprise AI and systems, and AI governance, with independent builds at the end.
[ Physical AI: industrial computer vision ]
Foreign material detection
Keeps contaminated product from leaving the line at a NASDAQ-listed pet food manufacturer, cutting recall exposure at the source. Real-time computer vision on live production cameras flags foreign material before packout.
LPN traceability
Gives the plant lot-level answers in minutes instead of shift-long searches, with every pallet traceable through production. License-plate-number tracking stitched into the line systems end to end.
Product cutting
Recovers yield lost to over-trim by guiding cuts to spec. Vision-guided measurement on the cutting line, tuned alongside the operators who run it.
Supported by anomaly detection, device health alerting, and an obstruction model, the workstreams that keep the line systems trustworthy.
[ Enterprise AI & systems ]
Information Hub / Piper
Made an internal assistant actually deployable: answers reach only the people cleared to see them. Led enterprise identity and permissioning (Entra ID, SharePoint) and media transcription, the foundation under the assistant, not the model.
Oracle blockchain engineering
Shared records institutions can trust without trusting each other. Blockchain software engineering on Oracle's enterprise ledger infrastructure.
Pro Drinks ERP
Runs a small manufacturer's whole operation in one system sized to the team that uses it: production, costing, pricing, shipments, and receivables in a single place instead of five spreadsheets.
[ AI governance & architecture ]
AI COE Portal
Lets every team ship AI with risk visible instead of guessed: one intake, one review path, and a Tier 1-4 risk framework that tells builders exactly what their use case requires before it ships.
Enterprise AI architecture
The foundation under a $1M+ AI program at a NASDAQ-listed pet food manufacturer: platform, patterns, and guardrails that let teams build without re-solving security and infrastructure every time.
[ Projects ]
- nawa: GTM production studio for tech companies across the US and MENA.
- SaniMektep: co-founded education initiative in Kyrgyzstan.
- Networking connectivity: systems build on network connectivity.
- Mini map: interactive mini-map build.
- Deep learning tool: tooling for training and inspecting deep learning models.
- Product management class project: end-to-end product spec and validation work.