AI & ML interests
Fine-tuned LLMs for embedded systems — ESP32, STM32, KiCad, SPICE, PlatformIO, TinyML, DSP, EMC. Building Kill_LIFE (spec-first agentic methodology) and Mascarade (multi-machine LLM orchestration). Open-source hardware meets AI.
Recent Activity
L'Électron Rare
"I would rather be a cyborg than a goddess." — Donna Haraway
Embedded systems engineering meets artificial intelligence. We build hybrid organisms at the intersection of silicon, language models, and live performance technology.
What we do
We fine-tune LLMs for hardware engineers — domain-specific models trained on ESP32, STM32, KiCad, SPICE, PlatformIO, DSP, and EMC documentation. Our models help embedded developers write firmware, design circuits, and debug hardware faster.
Open-source projects
- Mascarade — Multi-machine agentic LLM orchestration. P2P mesh, 8 providers, 29 models.
- Kill_LIFE — Spec-first agentic methodology for embedded systems development.
- KiC-AI — AI-powered PCB design assistant for KiCad.
- prima-cpp — Distributed LLM inference with pipelined-ring parallelism (CUDA + ZMQ).
Models & datasets
8 fine-tuned models and 9 curated datasets covering the embedded systems stack — from register-level STM32 programming to KiCad footprint generation and SPICE netlist analysis.
Links
The street finds its own uses for things. — W. Gibson