Gen AI Engineer at Skyovi · MS in Business Analytics and Artificial Intelligence, UT Dallas (Dec 2024)
I build things at the intersection of data and product. Right now that means GenAI systems at Skyovi. Before that it was ML pipelines, acoustic analysis tools, and econometric models. Most of my work is in Python and I care about making things that actually work in production, not just in a notebook.
SpeechMetrics · A vocal quality analyzer that takes an audio file and scores it across 10 dimensions: clarity, pace, engagement, expressiveness, and more. Built on librosa and Praat for acoustic feature extraction, with a FastAPI layer on top. I built this because every voice coaching tool I found was either a black-box SaaS or gave you one useless score. This one tells you exactly what is off and why.
Customer Churn Prediction · End-to-end pipeline that computes RFM features from retail transaction data and predicts churn using a Random Forest classifier. Covers the full lifecycle: feature engineering, model training, evaluation, and a clean setup for local reproduction.
Music and Mental Health Analysis · Econometric study on whether music listening habits correlate with self-reported mental health outcomes. Built in Jupyter with full EDA, statistical modelling, and a written report.
Core AI Models Journey · My scratchpad for implementing ML models from scratch. Linear regression up to neural nets. No frameworks, just math and numpy.
Python · SQL · R · LLMs · scikit-learn · FastAPI · librosa · Pandas · NumPy · Jupyter
I came into AI and data science through economics. My degree was heavy on econometrics and time series, which means I think carefully about causality and not just correlation. That background shapes how I approach every project whether it is a GenAI system or a churn model.
Open to connecting on interesting problems in AI and data.