Light Heart Labs | Infrastructure Contributors | Remote | Contract | Hourly We're building Dream Server (github.com/Light-Heart-Labs/DreamServer) — an open source, one-command bare-metal installer that deploys a full local AI stack. Think Linux of Local AI. Cross-platform, Apache 2.0, ~160 stars and growing. The mission: free and easy local AI for everyone. Nonprofits, small orgs, anyone who deserves sovereign AI infrastructure but can't afford Big AI dependency. Stack: Linux/Docker/systemd, vLLM, llama.cpp, Whisper, Piper TTS, Open WebUI, n8n, Qdrant, ComfyUI. Multi-GPU support shipping now. Looking for people who care about this problem — not just the work. Contributors from Palantir, LSE, and Teladoc already on the team. The last person we hired shipped a PR before asking for anything. Rates range with the work — $35/hr for hardening and reliability up to $100/hr for kernel-level and multi-GPU architecture. If you can do the work and want to help, we'll figure out a fair rate together. To apply: fork the repo, spend 30-60 minutes in the codebase, open a PR. Merged PRs get paid for the time spent building them. Bots need not apply. Long-term relationship for the right person.