Open to work

Computer Science @ Purdue 26

Shreya Komarabattini

Full-stack + AI software engineer turning complex systems into clear, useful products.

I build across frontend, backend, and applied machine learning, from real-time multiplayer experiences to NLP automation and data-driven tools.

Selected Work
2026

Moodle

Online drawing and guessing game you can play with friends. One person draws a secret word while the others try to guess it in the chat. You can make a room, invite others with a code, play against computer players, and even draw using hand gestures with your camera.

ReactSocket.IOPostgreSQLMediaPipe
2026

Protoplay

Work in Progress

Interactive platform that teaches engineering concepts through gamified simulations, architecture challenges, and algorithm puzzles by learning and building, not memorising.

Next.jsTypeScriptTailwind
2025–26

Traffic Sign Recognition

Senior Capstone team project focused on developing a mobile application that detects and classifies traffic signs in real time using a trained deep learning model. The app enhances driver safety by delivering fast and accurate traffic sign recognition through a live camera feed.

PythonDeep LearningComputer VisionMobile
2025

HealthyCal

Full-stack nutrition tracker with a React frontend, Node/Express API, and MongoDB storage. Helps users stay aware of how different foods contribute to daily energy and nutrient goals.

ReactNode.jsMongoDBVite
IT Ticket Routing Automation
2025

IT Ticket Routing Automation

End-to-end ML web app that reads free-text Helpdesk tickets and routes them to the right IT support group, predicting support group, issue type, and priority across 8 IT groups. Combines a trained NLP pipeline, a FastAPI backend, and a React dashboard with confidence scores and analytics.

PythonFastAPIReactscikit-learnXGBoostDocker
Fort Wayne, Indiana Crime Analysis
2023

Fort Wayne, Indiana Crime Analysis

Filtered 154,478 police activity rows into 30,336 likely crime incidents, then used Python to surface category, timing, reporting, and corridor-level patterns without overstating what the records prove.

PythonPandasKaggleData Viz
Categorical Data Visualization Study
2025

Categorical Data Visualization Study

Investigated how users interpret bar, line, and stacked bar charts. Designed user studies to evaluate visualization clarity, and formulated evidence-based guidelines for better data comprehension. Presented at Purdue's Annual Research Symposium.

ResearchUser StudyData Viz