Selected work

Streamlit app that turns messy retail operations data into clear waste signals — hotspots by stream, weekly trends, and downloadable summaries for the ops team.

Streamlit utility that fuses scattered CSV/Parquet drops into a single typed dataset — schema inference, dedupe, one-click export — plus a playground for forging quick data prototypes.

End-to-end MLOps pipeline for Marvel character data — ingestion, training, tracking, and deployment.

Experiments with declarative pipeline patterns on Apache Spark, captured in notebooks.

Databricks-native application exploring app patterns on the lakehouse platform.

Hands-on work with Databricks Unity Catalog — governance, lineage, and access patterns.

Running Kedro pipelines on Databricks — packaging, deployment, and orchestration notes.

Reference Kedro project demonstrating modular data engineering pipelines.

Notebooks exploring dimensional and analytical data modeling techniques.

Dockerised Apache Airflow setup for local development of orchestrated pipelines.

Interactive résumé built as a Streamlit app.

Notebooks demonstrating MLflow tracking, projects, and model registry.

Building web APIs with Flask for serving data and model endpoints.

Python chatbot experiment.

Streamlit application sandbox for rapid data app prototyping.

Collection of Streamlit experiments and component patterns.

Data engineering case study — modeling, transformations, and pipeline design.

Inventory management prototype exploring data flow and storage patterns.

Stocks data exploration under the bytemaster umbrella.

Stock market data experiments and analysis scripts.

MLOps testing ground — CI, packaging, and deployment experiments.

Source for this portfolio site.

Retail Store POS Streamlit app — Fuse & Forge: cashier, customer lookup, product catalog, tax, and checkout.

Neon Postgres setup powering the Retail Data Intelligence Ecosystem.

Declarative Lakeflow ingestion pipeline for the retail data ecosystem.