
Training frontier LLMs is brutally expensive — not just in dollars (clusters cost millions), but in time. Weeks or months on thousands of GPUs, with massive energy bills and idle hardware during synch
Applied AI Engineer || LLMs, Computer Vision, Backend Systems || Ex-ISRO • Ex-AegionDynamic • Stealth AI
AI Engineer who takes things from "cool experiment" to "actually running in production." I build with LLMs (RAG pipelines, chatbots), computer vision (YOLOv8, object tracking), and backend systems (FastAPI, databases). Sometimes I ship entire mobile apps in Flutter. What matters to me: does it work when real users touch it?
Markdown supported · ⌘↵ to post
No responses yet.