$ whoami --no-filter
> om gholwe 💀 — chronic shipper, recovering perfectionist
> 📍 india | 🌙 IST (Insomnia Standard Time)
> 🏆 5× hackathon winner | still loses arguments to semicolons 👀
> 🎯 mission: bullying small LLMs into running on potato phones 🥔
> 🧪 currently: making your phone smarter so the cloud stops snitching ☁️🚫
> 📊 100+ repos · ~8 actually slap · rest is "character development" 📈
> 📦 open-source contributor | professional issue creator & closer
> ☕ powered by: caffine and the fear of being mid
> 💌 dms open for: AI / mobile / edge-ML roles (and meme recs)
> ⚠️ side effects: may ship at 3am, refuses to use light modealso frying → website-builder · ClickClackArch · [ 91 more →]
┌─ on-device AI ─────────────────────────────────────────┐
│ • quantize big models → cram onto $100 android phones │
│ • multi-agent orchestration that runs LOCAL │
│ • LLM glue: gemma · qwen · deepseek wired into mobile │
└─────────────────────────────────────────────────────────┘
┌─ shipping ─────────────────────────────────────────────┐
│ • flutter front-to-back · idea → store in days not mos │
│ • prototype velocity > architecture astronautics │
└─────────────────────────────────────────────────────────┘
+ shipped a multi-agent LLM app that runs with airplane mode ON
+ got gemma running on a phone that struggles with whatsapp
- deleted node_modules to "free space" the night before a demo
- 3am commit msg: "it works i don't know why"
! 99 repos and counting · regret is a feature not a bugedge > cloud · ship > polish · vibes > specs · local > latency
dms open · @codecravings · let's build something that doesn't phone home 📵





