botboat.ai is a marine robotics & edge-AI consulting practice operating an instrumented 32-ft sailing testbed on San Francisco Bay. We architect ROS 2 / NVIDIA Jetson autonomy stacks, integrate NMEA 2000 sensor buses, and validate customer algorithms against RTK-GPS ground truth in live conditions.
Fixed-scope, fixed-fee engagements. Every engagement produces contemporaneous engineering logs, datasets, and a written report — deliverables you can put in front of your own stakeholders.
A 1983 Gulf 32 pilothouse sloop, refit as a floating robotics laboratory — berthed in Ventura with regular SF Bay operations. Real keel, real fouling, real EMI: the failure modes simulation never shows you.
| Hull | Gulf 32 pilothouse sloop, full keel, 12,500 lb displacement |
| Propulsion | Diesel auxiliary + autopilot-integrated helm actuation |
| Power | Solar array + AGM bank, instrumented DC bus telemetry |
| Operations | Ventura Harbor berth · San Francisco Bay trial range |
| Edge AI | NVIDIA Jetson class inference (30+ FPS vision pipelines) |
| Navigation | RTK GPS (cm-class), 9-DOF IMU (BNO085), fluxgate compass |
| Bus | NMEA 2000 backbone + NMEA 0183 bridges, full PGN logging |
| Hydrography | Water temp, salinity, depth; bilge & pump telemetry |
| Middleware | ROS 2, custom N2K gateways, ESP32 sensor nodes |
| Data | Timestamped mission logs, RTK ground-truth datasets |
| Benchmarks | Edge inference profiling on live camera + sensor feeds |
| Telemetry | The live board above streams from this stack's simulator when the vessel is not underway |
Businesslike from first contact: every engagement gets a numbered Scope of Work, a written deliverable, and an invoice trail.
Tell us about your platform and objectives. You'll get your numbered SOW instantly, and a reply from the principal engineer within one business day.