# CircuitGuard AI

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Facundo Gomez](https://cerebralvalley.ai/u/Procodice)
- **GitHub:** https://github.com/facunahuel420/CircuitGuard-AI
- **Demo video:** https://youtu.be/8P9SbIRwYW8
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3532

CircuitGuard AI v2.5 is a real-time, multimodal agentic diagnostic platform designed for microelectronics repair laboratories, service chains, and spare parts distributors. Built from five years of hands-on smartphone hardware repair experience and two years of AI automation engineering, it solves a massive industry bottleneck: thousands of motherboards, screens, and modules are discarded every day due to blind diagnostics, hours wasted tracing 10,000+ pad schematics pin by pin, and human errors on multimeter scales that permanently burn circuits.

CircuitGuard AI unifies three commercial modules (1. Hardware Lab, 2. Client Desk & Distributor Portal, and 3. Shop Owner ERP & Audit) and connects three synchronized real-time viewports:
- Visor A (Overhead Bench Camera): Monitors the physical motherboard (iPhone 11 820-01523-A) for spatial CAD alignment and optically audits the Elco 830D multimeter and USB tester—alerting technicians if the multimeter knob is OFF or set to the wrong voltage scale (-00.0) before measuring.
- Visor B (USB Optical Microscope + Voice/MP4 Video): Captures high-magnification SMD components, RF modules (AVAGO / Skyworks A1086-91), and BGA footprints with hands-free voice dictation and live 30s MP4 video recording.
- Visor C (CAD Vector Engine v2.5): Maps 10,210+ conductive pads, automatically illuminating active netlists (such as PP_VDD_MAIN across 175 pads) in Neon Red (#ef4444).

It also includes a Donor Board Viability Verdict to detect missing harvested ICs vs. intact reusable modules, a built-in Synth simulation mode so judges can test the entire platform without physical cameras, and a scalable architecture ready for home appliance inverter boards, industrial PCBs, and Meta Quest VR hardware.

## More from Meta Global AI Developer Hackathon

- [Open notebox](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3529)
- [StudBuds](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3530)
- [Dymes](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3531)
- [SokoSeal Team](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3533)
- [Saksi](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3534)
- [HiddenLot](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3535)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3532. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt