2+ years shipping production models at Devai Technologies. Classification, regression, anomaly detection at scale — with the MLOps to keep it running.
Hover or tap a node on the radar. These aren't aspirational — they're the skills I've shipped production code with.
Designed and deployed classification, regression, and ensemble models for behavioral prediction and anomaly detection across 20K–150K+ record batches. Ran 10+ experiment cycles, improving accuracy by 35% through iterative feature engineering and architecture refinement.
Built and maintained production data pipelines using Python, SQL, and Apache Spark to process large-scale datasets for training and evaluation — integrating new data sources and analyzing their downstream impact on model performance.
Deployed low-latency REST inference services (FastAPI) on AWS at <300ms, with experiment tracking, model monitoring, and drift detection to maintain stable production performance across evolving data distributions.
Collaborated with data science and engineering teams, documenting experiments, assumptions, and outcomes to maintain reproducibility — communicating technical findings to diverse stakeholders.
Three projects, three different problems, one engineer.
A rough simulation of the kind of ML inference system I build. Adjust the parameters, run a batch, and watch the latency distribution. Based on real patterns from my FastAPI + AWS deployments.
Simulated contribution heatmap across my last year of projects. Hover any cell.
I'm actively looking for ML Engineer roles — full-time, internship, or contract. Based in the United States, open to relocate.
This portfolio is fully keyboard-navigable. Try any of these: