the solo founder pipeline:
write the backend. train the model. build the API. dockerize it. deploy it. then open your inbox and do sales.
nobody told me the hardest part of running a studio would be context switching between pytorch and invoices.
built calendo from my room in lagos. no cofounder, no funding. just fastapi, supabase, and too much garri. it's live across 4 platforms now and real people use it. still feels unreal
quietly built a SaaS on the side. Calendo. schedule posts across 4 platforms from one calendar.
draft to published workflow, analytics, CSV export. FastAPI + Supabase + Next.js.
it's live. go try it right now.
Everybody thinks ML engineering is training models and watching accuracy go up. bro 80% of my time is cleaning data that looks like it was typed with elbows
things i've shipped as a backend & ML engineer:
- RAG pipeline for Supabase docs
- Calendo, a live SaaS across 4 platforms
- EB-1A petition analyzer
- SMS spam classifier w/ DistilBERT
just getting started.