Software Engineer | AI Systems & Product | Building Agentic AI that ships

🎓 B.S. Computer Science, B.S. Electrical Engineering, B.S. Innovative Leadership & Engineering Entrepreneurship
University of Illinois Urbana-Champaign (Grainger) · Class of 2029
📍 Chicago, IL

🔗 Portfolio: https://a2105z.github.io
🔗 LinkedIn: https://linkedin.com/in/aaravmittal

I'm a CS / EE / ILEE student at UIUC focused on agentic AI, AI infrastructure, and product — systems that reason, retrieve, and act, plus the full-stack products that put those capabilities in front of real users.

I've worked at Databricks and Rivian on data/AI platforms and production software. Alongside industry work, I contribute to open-source agent infrastructure around Databricks Omnigent and build AI-powered products end-to-end — from contract analysis and handwriting→LaTeX to scheduling systems and academic journey tools.

I care about reliable agent workflows, retrieval and memory-shaped features, clean infrastructure, and products people actually use.

• 🏗️ Experience on data / AI platforms at Databricks and production software at Rivian
• 🤖 Contribute to open-source agentic AI infrastructure (Databricks Omnigent meta-harness)
• 📱 Shipped CoParse — AI-powered contract analysis product (Android + FastAPI)
• ✍️ Shipped TeXForm — multimodal OCR product (handwriting → LaTeX) with a live demo
• 🧭 Built Meridian — full-stack product for academic journey tracking
• ⏱️ Built Chronos — constraint-aware time blocking / scheduling engine

• 🤖 Agentic AI
• 🧠 Long-Term Memory & Context Systems
• 🔍 Retrieval-Augmented Generation (RAG)
• 📊 LLM Evaluation & Reliability
• 🏗️ AI Infrastructure
• ⚙️ Distributed Systems
• 🎙️ Multimodal Interfaces (vision / OCR / document AI)
• 🚀 AI Product Engineering

Present

• Software / ML engineering on data and AI platform surfaces
• Working at the intersection of data infrastructure, ML workflows, and production systems that support AI applications

Present

• Software engineering in a high-reliability product environment
• Building software across systems and application layers for real-world products

Contract safety for students, renters, and early-career workers — an AI-powered Android product + API that highlights risky clauses, explains them in plain English, and suggests questions to ask before you sign.

• Shipped a full AI product loop: upload → extract → analyze → role-aware explanations → actionable questions
• Built Kotlin / Jetpack Compose Android client (Material 3) against a FastAPI + PostgreSQL analysis backend
• Designed async document jobs and structured analysis results for internships, leases, and freelance agreements
• Combined document AI with product UX: missing-protection checks and signature-readiness style scoring
• Monorepo with Docker Compose locally and Render + Neon deploy path

Stack

Kotlin • Jetpack Compose • FastAPI • PostgreSQL • Alembic • Docker • NLP / Document AI

🔗 GitHub: https://github.com/a2105z/CoParse

Handwritten notes → LaTeX. A multimodal OCR product: upload PDF/image, segment lines, run handwriting OCR, recognize math, and assemble downloadable .tex / PDF.

• End-to-end multimodal pipeline: segmentation → TrOCR → math recognition → LaTeX assembly
• Productized as React + FastAPI with Docker deploy and a public Hugging Face Spaces demo
• Optional MathPix / Pix2Text paths for formula recognition
• Designed for a real user workflow (notes in → usable LaTeX out), not just a model notebook

Stack

Python • FastAPI • React • TrOCR • Docker • Hugging Face Spaces • OCR / Vision-Language

🔗 Live Demo: https://huggingface.co/spaces/amittal417/texform
🔗 GitHub: https://github.com/a2105z/TeXForm

Academic journey tracker for high school and undergrad students — awards, scores, programs, goals — with accounts, filtering, analytics, and export.

• Product-shaped full-stack system: auth-scoped data, CRUD, search/filter, analytics, and export
• FastAPI + SQLAlchemy backend with relational modeling and SQL-backed reports
• React frontend for day-to-day tracking and dashboards
• Clear separation of product UX vs. legacy desktop prototype (Tkinter)

Stack

Python • FastAPI • SQLAlchemy • SQLite • React • Vite

🔗 GitHub: https://github.com/a2105z/Meridian

Intelligent constraint-aware time blocking engine — takes tasks, availability, and hard constraints, then builds a conflict-free calendar plan.

• Generates schedules with no overlap inside configured availability windows
• Applies hard constraints (protected blocks, max continuous work, priorities, deadlines)
• Supports preferred time-of-day preferences and .ics calendar export
• Full-stack product surface: React + TypeScript frontend, FastAPI backend, scheduling algorithms

Stack

Python • FastAPI • React • TypeScript • Algorithms • Calendar / Scheduling

🔗 GitHub: https://github.com/a2105z/Chronos

Active contributor / builder workflow around Databricks Omnigent — an open-source meta-harness for composing, controlling, and sharing AI agents.

• Agent composition across harnesses and models
• Policy / control at the meta-harness layer
• Reliable infrastructure for agentic products that need to ship

🔗 Omnigent: https://github.com/omnigent-ai/omnigent · https://omnigent.ai/

TI-84 inspired graphing calculator in the browser — Calculator, Graph, and AP Stats modes.

• Shipped a usable education product on GitHub Pages
• Multi-function graphing + stats workflows (LinReg, χ², distributions, plots)
• Optional FastAPI backend; client-side fallback for stats

🔗 Live: https://a2105z.github.io/Cosmos/ · GitHub: https://github.com/a2105z/Cosmos

Classic browser games (Chess, Tetris, Snake, Pong, 2048, …) — no accounts, just play.

• 14+ games with 2P / vs AI modes; Vite + GitHub Pages product surface

🔗 Live: https://a2105z.github.io/CoolGames.io/ · GitHub: https://github.com/a2105z/CoolGames.io

• 🤖 Agentic AI systems and Databricks Omnigent / meta-harness infrastructure
• 🧠 Memory, retrieval, and context patterns for AI products
• 🏗️ AI infrastructure and reliable ML/data platforms
• 🚀 Shipping AI-powered products end-to-end (mobile, web, multimodal)
• 📊 Evaluation and reliability habits for agent / LLM features

I'm currently seeking Software Engineering, AI Engineering, and AI Product internships, particularly in:

• Agentic AI
• AI Infrastructure
• Machine Learning Systems
• Retrieval / Memory Systems
• LLM Platforms
• Distributed Systems
• AI Product Engineering

If you're building in these areas, I'd love to connect!

📧 Email: aarav.mittal.ma@gmail.com

🌐 Portfolio: https://a2105z.github.io

💼 LinkedIn: https://www.linkedin.com/in/aaravmittal

💻 GitHub: https://github.com/a2105z