Fullstack developer. Building web apps by day, exploring local AI inference & biological computing by night. 🧠🤖

Day job: Fullstack web development

• Frontend: Angular, Qwik, React, WebAssembly
• Backend: Fastify, Node.js
• Database: PostgreSQL (Supabase), MongoDB, Oracle

Side projects & Open Source:

• 🧬 Cortical-DAC Convert: Developed a client-side WebAssembly tool for the Cortical Labs ecosystem to process and convert massive .h5 biological computing recording files (>2GB) directly in the browser.
• 🤖 DS4 Italian Documentation: Built the reference documentation and guide for antirez/ds4, covering local inference setups, 2-bit quantization, and distributed execution.
• 🛠️ Local AI Systems: Testing hardware boundaries by exploring what it takes to run frontier LLMs locally on consumer hardware. Forked ds4 to experiment with AMD HIP/ROCm porting.

• AMD Ryzen AI Max+ 395 (Strix Point)
  • 16 cores / 32 threads
  • 128 GB RAM
  • Radeon 8060S (62 GB VRAM)

• Biological Computing: Silicon-neuron interfaces and the Cortical Labs ecosystem.
• Local LLM Inference: Running DeepSeek V4 Flash with IQ2 quantization (no cloud dependencies).
• Hardware Acceleration: AMD GPU computing via ROCm/HIP and client-side optimization via WebAssembly.
• Edge AI: Building functional, private, and on-device applications (like custom LM Studio integrations).

• LinkedIn
• Blog