Prasanth Muntha | AI/ML Engineer

01 / About

Building systems that think

Currently pursuing my MS in Artificial Intelligence at Yeshiva University in New York. I specialize in bridging research and production — turning papers into pipelines.

My work spans local-first AI agents, 3D computer vision, and cloud-native data platforms. I build things that are fast, private, and actually work in the real world.

02 / Stack

Tools I use

Languages

Python TypeScript JavaScript SQL Bash

AI / ML

PyTorch LangChain Vercel AI SDK Ollama ONNX HuggingFace Scikit-learn

3D / Vision

PyVista VTK OpenCV MoveNet YOLOv8

Data

Pandas Airflow ChromaDB PostgreSQL

Frontend / Full-stack

Nuxt Vue Tailwind FastAPI

Cloud / DevOps

AWS Vercel Docker Git

03 / Projects

Things I've built

01 Latest Try it live

Glacix

AI Chatbot

Full-stack AI chatbot with Nuxt 4 and Vercel AI SDK. Multi-model support (NVIDIA NIM), RAG over documents (PDF, CSV, Excel, Word), live web search with citations, and per-user token usage tracking in PostgreSQL.

Nuxt 4 Vue 3 TypeScript Vercel AI SDK PostgreSQL Drizzle

githubx

AI that converts complex GitHub READMEs into simple, OS-specific installation guides. No more copy-pasting broken commands.

FastAPI Ollama Gemma

Local-First Agentic RAG

Privacy-first AI with "Flight Mode" — air-gapped inference, zero data leakage. Routes queries between Vector Search, Web, and Python Analyst.

LangChain Llama 3.2 ChromaDB

3D Scoliosis Detection

Production 3D AI pipeline: raw body scans → mesh rectification → pose detection → ROI segmentation → Cobb angle. 99% view classification.

PyVista MoveNet ONNX

Fake News Detection

Hybrid verification: BERT classification + LLM semantic fact-checking. Explains why a claim is unreliable, not just that it is.

BERT Llama 3 NLP

04 / Contact

Let's talk

Looking for Summer 2026 internships in AI/ML, 3D Vision, and Data Engineering. If you're building something ambitious — I want in.

Email LinkedIn GitHub