# Bruno Seilliebert

> Bruno Seilliebert, Software Engineer. I build systems end-to-end, from the data pipeline to the pixel.

- Title: Software Engineer
- Location: Paris
- Email: bruno@seilliebert.dev
- GitHub: https://github.com/bruno00o
- LinkedIn: https://www.linkedin.com/in/bruno-seilliebert/
- CV: https://bruno.seilliebert.dev/cv.md
- CV (PDF): https://bruno.seilliebert.dev/cv.pdf
- Canonical: https://bruno.seilliebert.dev/

---

I enjoy building systems end-to-end, from the data to the display. I'm currently helping build a real-time digital twin of the SFR network.

## Work

- [Kazimo](https://bruno.seilliebert.dev/work/kazimo.md): An open source connected object for an elderly person living alone: photo frame, video calls, messages and a voice assistant that can show things on screen. No apps, no menus.
- [Aroma](https://bruno.seilliebert.dev/work/aroma.md): A web app to check your university schedule on mobile, where ADE (the platform used by several French universities) is barely usable on a phone. Current class and room visible the moment you open it.
- [GameRoute](https://bruno.seilliebert.dev/work/gameroute.md): A Windows app that automatically identifies the servers of running games and traces the network path to them.
- [Homelab](https://bruno.seilliebert.dev/work/homelab.md): A self-hosted Kubernetes cluster running my apps, my personal services and the connected home.
- [claude-dial](https://bruno.seilliebert.dev/work/claude-dial.md): A desk companion for Claude Code: a round screen that shows what each session is doing, and a dial to approve or reject a permission without leaving the terminal.
- [Jarvis](https://bruno.seilliebert.dev/work/jarvis.md): A local voice assistant for the home, inspired by Tony Stark's Jarvis (as any self-respecting assistant should be), wired into my homelab services.
- [Archive](https://bruno.seilliebert.dev/work/archive.md): Older or smaller projects, kept here for the record.

## Stack

- Languages: TypeScript, Java, Scala, Kotlin, C/C++, Python, PHP, Go, Rust
- Data: Spark, Kafka, SQL, Neo4j, ClickHouse, MongoDB, Redis
- Infra: Docker, Kubernetes, Talos
- Frameworks: Next.js, Tanstack Start, Java Spring, Hono, Tauri
- CI/CD: GitHub Actions, GitLab CI
- AI: Claude, MCP, Ollama, LM Studio

## Experience

- 2025 → present: Software Engineer, SFR (Full-time)
- 2023 → 2025: Software Engineer, SFR (Apprenticeship)
- 2022 → 2023: Web & Mobile Developer, Kynova (Apprenticeship)

## Education

- 2023 → 2025: Master's, Software & Data Engineering, Université Gustave Eiffel (Apprenticeship)
- 2020 → 2023: Bachelor's, Computer Science, Université Gustave Eiffel

## Now

- building: Kazimo, Jarvis, GameRoute.
- exploring: Local LLMs.
- running: A homelab Kubernetes cluster named Going Merry.

## Notes

- [Kazimo runs on Mistral, and could run at home](https://bruno.seilliebert.dev/writing/kazimo-mistral-local.md): Kazimo's whole voice chain, transcription, agent and synthesis, goes through Mistral models with open weights. Why I want to run it locally, and what machine it would take, in three estimated tiers.
- [Claude meets Gaspard and helps me remember 5 years of CS classes](https://bruno.seilliebert.dev/writing/gaspard-rag-mcp.md): An MCP server plugged into a LightRAG index (graph + vectors) over 5 years of class material, callable straight from claude.ai. All local on the Mac and the Talos cluster, zero tokens billed server-side.