William Elimbi

I am a senior full-stack developer specializing in 100% self-hosted, privacy-first local AI solutions. By 2026, companies and creators will increasingly reject OpenAI/Groq/Anthropic APIs due to costs, data leaks, and GDPR. I build high-performance RAG (Retrieval-Augmented Generation) systems for them, deployed on their own servers/VPS, without any external cloud dependency.

My core strengths:

Self-hosted RAG: document/PDF/website ingestion, local vector DB (Chroma/Qdrant/FAISS), efficient retrieval + generation with open models (Llama 3.1/Phi-3/Mistral/Nemotron via Ollama or llama.cpp). One-click Docker Compose deployment, Bun.js for ultra-fast backend.

→ Concrete example: GitHub repo with full local RAG in Bun.js (scrap + chat interface) → github.com/sepiropht/[your-rag-repo]

Video subtitle automation: Whisper pipeline + post-processing (FR/EN correction, precise timing, SRT overlay → MP4). Ideal for YouTube, podcasts, training, Reels/TikTok. Batch or one-shot processing, optimized performance on GPU/CPU.

→ Demo repo: github.com/sepiropht/auto-subtitles (Whisper + ffmpeg/Bun)

Custom Rust/Bun tools: High-performance CLIs, respectful scrapers, Rust Git-like implementations (rgit), Node → Bun migrations for 2-3x speed. OVHcloud production experience (cloud/DevOps).

Full-stack: React/Next.js frontend if needed, REST/GraphQL APIs, secure auth.

Why me?

Full-time availability.

Pragmatic approach: I deliver clean, testable, documented code, with Docker + basic CI/CD included.

Realistic freelance rate 2026: €550–€700 HT / day (adjustable for short missions or packages). Typical missions: 3–15 days for an MVP/RAG, or longer for full integration.