I am a Machine Learning Engineer and Full-Stack Developer currently pursuing my M.S. in CS at Georgia Tech. I specialize in bridging the gap between state-of-the-art research (CV, Deep Learning) and production-grade software (Distributed Systems, Cloud Infra).
- 🔭 Research: Published in Nature Scientific Reports (ML for Marine Conservation).
- 💼 Experience: Scaled chat systems to 100k+ users at Visible; improved solar forecasting by 39% at Rutgers.
- 🎓 Academics: Rutgers University CS & Data Science (Summa Cum Laude, 4.0 GPA).
- 🏆 Awards: Matthew Leydt Society (Top 2% of Rutgers Graduates), 1st Place Rutgers Data League.
| Domain | Technologies |
|---|---|
| Languages | |
| ML & Data | |
| Web & Backend | |
| DevOps & DBs |  |
Concepts: Data Structures & Algorithms • System Design • OOP • Cloud Computing • Unit Testing • Agile
|
A real-time energy dashboard forecasting grid carbon intensity and recommending EV charging windows.
|
🎼 DeepJAn intelligent DJ application that reads the room's energy to curate the perfect musical vibe in real-time.
|
|
Full-stack degree planning platform for Rutgers with an integrated AI assistant.
|
Deep learning pipeline converting grayscale images to realistic color using CIELAB color space.
|
|
Predictive model for U.S. college tuition costs based on socio-economic indicators.
|
End-to-end autonomous agent that navigates real-time gameplay loops using computer vision.
|
My research focuses on applying Machine Learning to environmental and humanitarian challenges.
🐋 Nature Scientific Reports (2024)
Machine learning for modeling North Atlantic right whale presence
Co-authored a paper improving whale detection recall from 16% → 56% using ensemble learning on satellite and glider data.
Read the Paper
🗺️ Refugee Settlement Detection
Automated refugee settlement detection via Python ML pipelines using Google Cloud and satellite imagery.
Achieved 82% classification accuracy, significantly improving scalability over manual mapping efforts.
🌞 Solar Forecasting
Built a CNN pipeline using sky image sequences to improve short-term solar irradiance forecasting accuracy by 20–39%, aiding in renewable energy grid stability.