I design and ship end‑to‑end AI systems and data platforms—from experimentation to production. My focus areas include foundation model fine‑tuning, multimodal perception, realtime ML on streams, and scalable analytics for business and quantitative research.
- MLOps: experiment tracking, CI/CD for ML, feature stores, model registry, evaluation at scale
- Applied ML: NLP, vision, time‑series forecasting, tabular ML, recommender systems
- Systems: distributed compute, streaming, vector databases, GPU acceleration, low‑latency infra
- EduScore Predictor — Feature engineering + linear regression on real exam data. Issue • Repo
- AgriYield Insights — Crop yield forecasting with regression baselines and error analysis. Issue
- SalesFlow Forecaster — Time‑series forecasting for sales; seasonality + promotion effects. Issue
- PropVal AI Engine — Real‑estate price prediction with model comparison. Issue
These projects highlight applied ML workflows. Notebooks are stored in each repo under /notebooks.
- Binary-Intelligence-Frameworks-Logistic-Regression- — End‑to‑end classification with metrics and calibration.
- Predictive-Intelligence-Systems-Linear-Regression- — Feature pipelines, cross‑validation, and diagnostics.
- Market-Alpha-Discovery — Alpha research notebooks and experiments.
- microsoft/BitNet — Official inference framework for 1‑bit LLMs.
- huggingface/course — Learn transformers and modern NLP.
- duckdb/duckdb — In‑process analytics DB for notebooks and apps.
- Docs and feature contributions across ML repos; currently exploring issues in Infisical and AWS SDK v2.
Email • LinkedIn • Hugging Face