Atlanta, Georgia

Aaryan Kapoor

Advanced
@Aaryan-Kapoor

AI Engineer & Researcher | LLMs, Agents, Inference Optimization | 10k+ weekly downloads | 1000+ GPU hrs/mo | Multi Hackathon Winner | Open to work :)

z-image-turbo. A professional web interface for the Tongyi-MAI Z-Image-Turbo model — lightning-fast text-to-image generation with 6B parameters.

135

ModelGate-Hackathon. 🏆Winning Project | ModelGate is a contract-aware AI control plane that ingests customer contracts, extracts SLA/privacy/routing constraints, and generates an OpenAI-compatible endpoint that automatically routes every request to the optimal model. Simple queries go to cheap models. Complex queries go to premium ones.

50

24hr-research-agent. An experimental autonomous research system that conducts comprehensive, multi-hour research sessions and produces book-length reports with full citations on any topic.

46

NvidiaOmniEmbed. A local, multimodal embedding server and vector database using NVIDIA's Omni-Embed-Nemotron-3B. This project allows you to generate embeddings for Text, Images, Audio, and Video on consumer hardware (tested on RTX 5070 Ti / 12GB VRAM), save them to a local index, and perform multimodal semantic search.

11

simple-ntfy-mcp. A simple MCP (Model Context Protocol) server for sending push notifications via ntfy.sh. Works seamlessly with Claude Desktop and other MCP-compatible clients.

7

Reflect-KSU-Hackathon. 🏆Winning Project | Reflect is a cutting-edge mental health counseling platform designed to bridge the gap between students and counselors. By integrating real-time video therapy with advanced AI analysis, mood tracking, and personalized care plans, Reflect provides a holistic approach to mental well-being.

5

llama.cpp. LLM inference in C/C++

5

mcp-character-tools. The last thing you need for your LLM to work with individual characters or count the number of r's in a word. This is an MCP server providing 14+ comprehensive (and pretty) character and text analysis tools to help LLMs work with individual characters - something they struggle with due to tokenization.

3

prims-algorithm. JavaScript

1

SOTA-coder. An advanced AI coding agent with a custom runtime, tool execution, MCP support, and a beautiful web interface.

1

genesis-synthetic-datasets. Genesis is an autonomous agentic pipeline that generates high-quality synthetic datasets for machine learning. Simply describe what you need, and Genesis handles research, prototyping, alignment, and mass production.

1

ai-dynamic-exams. An experimental Local-first adaptive exam system for universities, powered by retrieval (RAG) + an OpenAI-compatible LLM endpoint. Please note that this project is intended as a proof-of-concept and needs additional work to get it to the MVP stage :)

1

MassEmailLookup. This repository contains an Email Lookup Tool that leverages the Perplexity API to retrieve contact information such as email addresses, phone numbers, and sources for individuals from an Excel file. The GUI for the tool is built using PySide6 for a modern, frameless interface with dark mode support, and it includes API key storage encryption.

1

vector-search-presentation. HTML

1

d2l-cli. Read-only CLI for D2L Brightspace. Pulls grades, assignments, content, syllabi, and more - designed to be used by AI coding agents (Claude Code, OpenClaw, etc.) as a tool.

1

Surface. TypeScript

1

degreeworks-cli. Read-only DegreeWorks CLI for KSU – built for AI agents to plan student schedules

1

simple-brightness-tray. PowerShell

1
18
Apply