Ramsaheb Prasad

I’m an AI/ML Engineer focused on Generative AI and LLM systems, with hands-on experience building end-to-end AI backends and reliable AI workflows. I enjoy turning complex model behavior into low-latency, production-ready experiences.

In my AI Developer Intern role at Arkham Archives (Remote), I built a scalable LLM-powered backend using FastAPI and Supabase to enable adaptive question generation, automated evaluation, and real-time student analytics. I also reduced manual effort by approximately 60% by delivering technical interview and coding practice modules that performed with low latency.

I’ve worked across the stack as well—integrating the backend with a large-scale React frontend and collaborating on GCP deployment for complex APIs and real-time data flow. I’ve implemented Generative AI APIs (text-to-image, music, video) and built a RAG-based chatbot for personalized learning support.

My project work reflects a strong bias toward measurable reliability: I led an AI-Powered Classroom Attendance & Engagement System using YOLOv8, MTCNN, MediaPipe, and FaceNet, and I built an AI Output Verification Platform to detect hallucinations using embedding similarity and NLI with audit logging for ALLOW/FLAG/REFUSE decisions.