Real systems. Shipped.

Problem

Riders had no objective way to measure pressure distribution and balance during training. Coaches relied on intuition.

System architecture

• Custom multi-sensor PCB embedded in a saddle pad
• ESP32 firmware with BLE streaming + onboard buffering
• Flutter mobile app with live + post-session analytics
• Cloud sync, user accounts, session history

Outcome

Funded venture with award recognition. Working hardware, app, and brand — built end-to-end as founding engineer.

Problem

Whitewater Brewing needed reliable, auditable, real-time monitoring of process sensors to protect quality and support compliance in a live production environment.

System architecture

• Sensor acquisition + edge processing
• PostgreSQL time-series storage
• REST backend with auth + audit log
• Operator dashboard with live + historical views

Outcome

Deployed to production at Whitewater Brewing. Day-to-day reliability for a working facility, with audit-ready records and live operational visibility.

Problem

During industry work at Analog Devices, reference designs and demo platforms needed to be redesigned, debugged, calibrated, and made stage-ready for trade shows.

System architecture

• Schematic + layout review and revision
• Sensor-system bring-up, calibration, and test
• Hardware debug and signal-integrity investigation
• Demo-platform software and firmware support

Outcome

Demo platforms and sensor systems improved for real event use, with testing and calibration work focused on system performance.

Problem

Teams want long-context, private LLM capability without sending data to third-party APIs — but most setups don't actually work in production.

System architecture

• Multi-GPU server build, drivers, thermals
• Local inference stack (vLLM / llama.cpp class)
• RAG pipeline with vector store + reranking
• Internal API surface for app integration

Outcome

Working local inference for large-context workloads. Real engineering — not hype, not a demo.