AI-powered electronics ecosystem combining hardware shopping, guided learning, AI schematic assistant & community forum.
Beginners and engineering students struggle when debugging hardware circuit diagrams, choosing compatible components (e.g. voltage matching between ESP32 and 5V relays), and sourcing verified components across fragmented vendors.
CircuitBotz unifies hardware learning with an intelligent circuit assistant. Users can upload schematics or prompt with component lists to receive automated pinout checks, power consumption estimations, and direct checkout for needed resistors, sensors, and boards.
Client-server architecture with React frontend interfacing with an Express REST API. The AI assistant leverages streaming LLM responses with grounded hardware system prompts and domain-specific circuit verification rules.
Responsive single-page application with real-time schematic rendering and e-commerce cart management.
Processes prompt queries, validates pin configurations, and serves catalog queries with sub-50ms latency.
Document database storing product specs, user project guides, and cached hardware pinout diagrams.
Standard LLMs frequently hallucinate GPIO pins on microcontrollers like ESP32-WROOM-32.
Implemented a strict JSON grounding schema paired with a deterministic hardware validation lookup table that verifies GPIO restrictions before returning the schematic to the user.
High concurrency during flash sales caused cart race conditions.
Engineered optimistic UI updates on the client combined with atomic database reservations in MongoDB.
500+ active student builders tested the platform during university tech events
Over 1,200 circuit queries successfully resolved with 94% accuracy score
Reduced BOM assembly time for student robotics teams by 40%