Trends change fast. Durable systems never go out of style.
nicholas cambre · systems & automation engineer · seattle, wa
My work is about stripping away excess and building practical foundations that stand the test of time. Whether it's an AI assistant, a smart-home automation flow, or a streamlined server, the goal is the same: technology that feels effortless to use, resilient under pressure, and meaningful in the real world.
Featured projects
Jemma AI
A local-first AI assistant built from scratch — C++ inference, streaming voice, real auth, and tool calling that runs on hardware I own.
Jemma is a locally hosted conversational AI assistant with memory, streaming voice, and infrastructure integration — designed to be useful, observable, and entirely independent of cloud-based assistants.
Jemma Voice Pipeline
Local neural text-to-speech that gives Jemma her voice — 23ms per sentence on a 2080 Ti. GLaDOS-inspired, never leaves the network.
A self-hosted Python TTS service that turns Jemma's responses into audio in real time. 23ms per-sentence inference on consumer GPU hardware. Inspired by GLaDOS from Portal 2 because every other LLM sounds the same.
Stack
- TypeScript
- Next.js (App Router)
- React
- Python
- Linux
- Home Assistant
- LLMs & Orchestration
- MQTT
- Docker
- Nginx
- Cloudflare
- Git
- Automation
- DX
- Resilient Systems
- TypeScript
- Next.js (App Router)
- React
- Python
- Linux
- Home Assistant
- LLMs & Orchestration
- MQTT
- Docker
- Nginx
- Cloudflare
- Git
- Automation
- DX
- Resilient Systems