With a background in AI + Chemical Engineering (CMU), I focus on:
- Scalable data pipelines (PySpark, distributed systems)
- Deep learning & time-series modeling (PyTorch)
- End-to-end ML systems (data → model → deployment)
- High-performance computing (HPC, large-scale training)
I enjoy working at the intersection of ML + systems + real-world impact — whether it's anomaly detection in industrial processes, fraud detection, or building intelligent pipelines that scale.
Currently seeking full-time ML / AI Engineer roles where I can build, optimize, and deploy impactful systems.
-
Agentic Financial Fraud Detector
Built a multi-model fraud detection system (XGBoost, RF, AE, IF) with SHAP interpretability, LLM-based explanations, and adversarial robustness evaluation via noise and camouflage attacks. -
Time Series Anomaly Detection (CMU Research)
Co-authored a recently released preprint benchmarking ML fault detection on Tennessee Eastman Process dataset, designed, implemented and evaluated an LSTM-FCN model with best test-set performance (99.37% test accuracy) among all benchmarked models across 18 fault classes and millions of time-series samples. -
Spotify to MP3 Converter - This is a Spotify to MP3 converter that allows users to easily convert Spotify playlists into downloadable MP3 files. The app processes Spotify playlist URLs and provides a simple way to download each track via links that redirect to the corresponding MP3 files hosted on the server.
Languages & Data:
Python · SQL · PySpark
ML / DL:
PyTorch · TensorFlow · Scikit-Learn