Notionist iconMG

Manit Garg

I design and build performant, human‑centered software. My work spans scalable web platforms, applied machine learning, and developer tooling. I care about clean architecture, delightful UX, and shipping quickly without sacrificing quality.

👇 Scroll down to explore
ResumeGitHubLinkedInContact

About

I'm Manit Garg, a Computer Science student at Penn State, pursuing my B.S. with a Business Fundamentals Certificate. My academic background covers data structures & algorithms, machine learning, database systems, and software engineering, paired with strong experience in AI research and full-stack development.

My technical expertise spans Python, TypeScript, JavaScript, Java, C/C++, SQL, and frameworks like Next.js, React, Node.js, TensorFlow, and PyTorch, with a strong focus on scalable architectures, data pipelines, and AI/ML applications. I'm passionate about building innovative, high-impact solutions that bridge cutting-edge research with practical use cases.

Research

Explainable AI Research

2025 — Present

Pennsylvania State University

Developing explainable AI models for program semantic equivalence, applying graph-based learning techniques (GCN, GAT) and embedding methods like Word2Vec, Code2Vec, and DeepWalk.

Achieved 92% accuracy in detecting obfuscated code similarities, supporting advancements in automated code review and verification tools.

Contributing to research that bridges the gap between AI interpretability and practical software engineering applications.

Projects

Buddy — AI-Powered Penn State Assistant

Next.js, TypeScript, Python, LLMs, PostgreSQL • Sept. 2025 — Present

  • Architecting AI chatbot at chat.mlpsu.org delivering course planning, advising, and Penn State resources
  • Engineering multi-source data pipeline spanning PSU Bulletin, department handbooks, and RateMyProfessor data
  • Building prerequisite validation engine that handles complex requirement chains and general education rules
  • Implementing citation and provenance tracking to surface rule sources with freshness metadata
  • Developing OCR + NLP transcript parsing pipeline with confidence scoring and user confirmation flows
  • Designing privacy-first architecture with encrypted grade storage and instant data deletion controls
Visit Buddy

Mimicry — Real-Time AI Voice Detection Game

Next.js, TypeScript, LiveKit, ElevenLabs, Deepgram • Oct. 2025

  • Engineered real-time voice game training users to detect AI voices through competitive gameplay
  • Integrated LiveKit for ultra-low-latency audio, plus ElevenLabs for instant 10-second voice cloning
  • Coordinated LiveKit, ElevenLabs, Deepgram, Claude, Groq, and Supabase in a real-time transcription loop
  • Built matchmaking system with private lobbies, QR sharing, and post-game analytics via Recharts
  • Launched educational platform at play.mimicry.fun with player profiles and progression tracking
Play Mimicry

StudySprint — AI-Powered Coding Platform

Next.js, TypeScript, Python, PostgreSQL, Supabase • June 2025

  • Architected platform combining 3,600+ LeetCode problems with AI-driven learning analytics and search
  • Built automated Python ingestion to fetch, transform, and sync structured problem data from multiple APIs daily
  • Implemented intelligent search with real-time filtering for rapid discovery of relevant coding challenges
  • Developed responsive TypeScript frontend with Monaco Editor and optimized PostgreSQL queries
  • Designed scalable database with 5 interconnected tables powering real-time progress tracking and analytics
Visit StudySprint
Game Captcha — AI-Native Bot Detection

Game Captcha — AI-Native Bot Detection

Next.js, React, TensorFlow, Deep Q-Networks • Apr. 2025

  • Created CAPTCHA alternative marrying interactive gaming with AI behavioral analysis for bot detection
  • Benchmarked with pre-trained Deep Q-Network to separate human and automated gameplay patterns
  • Shipped modular, embeddable verification system backed by scalable infrastructure on Railway
  • Implemented WebXR compatibility for immersive verification with 95%+ human detection accuracy
  • Won Timeless Tech Track at HackPSU Spring 2025 among 300+ projects
Visit Game Captcha
Swipe — AI-Driven Professional Networking

Swipe — AI-Driven Professional Networking

React, Node.js, MongoDB, OpenAI API • Oct. 2024

  • Built networking platform that leverages LLMs to match professionals on research interests and goals
  • Implemented NLP pipelines to analyze profiles and compute semantic similarity scores for matching
  • Designed swipe-based interface with real-time messaging to accelerate professional introductions
  • Scaled MongoDB data layer powering algorithms that process over 1,000 user profiles
Project links coming soon

Leadership & Clubs

Treasurer & Event Coordinator

2024 — Present

Machine Learning @ Penn State

• Leading budgeting for 200+ members and securing over $15K in funding

• Organizing technical workshops and networking events with industry professionals

• Mentoring peers in machine learning concepts and practical applications

Contact

Open to collaboration, internships, and impactful product work. Let's build something amazing together!

📧 Email Me📄 Resume