A Next.js application that helps users create personalized learning plans using the Vercel AI SDK. Supports OpenAI and Cerebras inference backends.
Biological
Coding Agent
I architect end-to-end product systems that blend AI, orchestration, and full-stack engineering, turning ambiguous ideas into resilient, scalable software.
The agent that listens to the internet and just… fixes your code. Auto-groups feedback, spins up instances, writes code, and opens PRs automatically. Built at the AIE Code Agents Hackathon.
A modern portfolio website showcasing my work as a software engineer, built with Next.js 16 and featuring a futuristic cyberpunk aesthetic with animated backgrounds and smooth interactions.
FastAPI + LangGraph service that extracts submittal requirements from specs and verifies submittal packages using an LLM + OCR pipeline, with optional Procore automation.
A cross-platform 3D product visualization experience for e-commerce, streaming neural-rendered product views from an AWS-powered pipeline.
CME 295: Transformers & LLMs
Stanford course exploring Transformers and Large Language Models (LLMs), covering architecture, NLP evolution, and real-world applications.
Context Engineering Intro
Context engineering is the new vibe coding - it's the way to actually make AI coding assistants work.
The Odin Project
A full stack curriculum that is free and supported by a passionate open source community.
Aura Design Assistant
Documentation for Aura, an AI-powered design assistant for creating beautiful designs with ease.

Naga Karumuri
Software Engineer
Building at the intersection of AI and Product.
I build AI systems that amplify what humans can do. My projects range from autonomous agents that monitor community feedback and open PRs, to LLM pipelines that verify construction compliance, to neural rendering systems for product visualization. I work across the stack because shipping requires understanding the whole system.
I've built with FastAPI, Next.js, LangGraph, Flutter, and trained models on H100 clusters. I ship quickly, learn from real usage, and iterate. The goal is working software that solves real problems, not perfect architecture that never ships.
Full Stack
End-to-end development with modern frameworks.
AI-Powered
Integrating LLMs for intelligent applications.
Scalable
Robust architecture designed for growth.