Zero to 12 Production AI Projects in 12 Months
Independent Consulting
Situation
In January 2025, I had zero AI/ML experience. My entire career was wireless networking — 25 years of RF, enterprise infrastructure, and protocol debugging. But I saw the AI wave transforming every industry and knew that my networking domain expertise combined with AI capabilities could create something unique. I decided to go all in.
Task
Teach myself AI engineering from scratch — not just prompting, but building production applications: LLM orchestration, computer vision, vector databases, full-stack AI-powered SaaS. Do it fast enough to be relevant, and ship real products, not tutorials.
Action
I immersed myself completely. Built Switchboard (LLM routing across Claude, GPT-4, Gemini with 83+ agent profiles) to understand multi-model orchestration. Created Claude Command Suite (148+ commands, 54 AI agents) as my daily development tool. Shipped Skyforge with computer vision (YOLOv8, CLIP) for automated drone footage analysis. Architected FlightDeck as a full-stack SaaS (FastAPI + Next.js + PostgreSQL/pgvector + Stripe). Built an ACS Meeting Assistant with WhatsApp/Twilio/ElevenLabs integration. Each project forced me deeper into the stack — from prompt engineering to vector embeddings to real-time streaming.
Result
Shipped 12+ production applications in 12 months. Built a genuine AI engineering skill set spanning LLM orchestration, computer vision, RAG pipelines, agent systems, and full-stack AI-powered products. Founded AI Aerial Solutions as a vehicle for drone + AI work. The speed of learning demonstrates exactly the adaptability that matters in a fast-moving field — I do not wait for training courses, I build.