Who is George?

I’m a data analytics engineer focused on building end-to-end data products: from ingestion and transformation, through modelling and analysis, to clear outputs that teams can actually use. I enjoy working at the intersection of data engineering, data science and analytics, where the goal is not just to build something clever, but to deliver something reliable and measurable.

In my current role at Ishango I work with multi-million row datasets across several clients, rebuilding and optimising ETL pipelines using tools such as SQLAlchemy, Pydantic, DuckDB, & Parquet files, and deploying them on Azure. I design and maintain Azure Functions and database logic that keep data flows robust and performant, and I integrate external systems via APIs to keep operational data in sync and reduce manual work.

On the analytics and BI side, I build Power BI dashboards and reporting layers that support day-to-day decision making for internal teams and clients. I expose analytical logic through FastAPI endpoints and SQL workflows, so that insights and metrics can be reused across applications instead of being locked into a single report. I’m comfortable owning the full lifecycle of these solutions: scoping the problem, designing the data model, implementing, testing and iterating based on feedback.

My technical toolkit includes Python (pandas, NumPy, scikit-learn, TensorFlow, PyTorch, Matplotlib, NLTK, spaCy, Selenium, BeautifulSoup, SQLAlchemy, Pydantic, FastAPI), SQL (PostgreSQL, MySQL), R and SPSS, as well as version control, data visualisation libraries (Seaborn, Plotly, Matplotlib) and modern cloud tooling (Microsoft Azure). I have experience working with APIs, scheduling and automation, and I care a lot about making pipelines observable and maintainable rather than just “working once”.

Beyond pure engineering, I’ve worked on a range of machine learning and analytics projects: time-series forecasting for wind turbine health, autoencoder-based fraud detection, hotel booking cancellation prediction and guest segmentation, customer RFM analysis, an ML-driven R Shiny app for drug classification, and network analytics to study how AI knowledge spreads on a trading floor. These projects span classic supervised learning, unsupervised clustering, deep learning and NLP, and they all share the same goal of turning complex data into practical insights.

I enjoy collaborating with product, business and technical stakeholders, translating questions into well-defined analytical problems and then into concrete solutions. Whether that means re-architecting a pipeline, designing a model or mentoring a junior analyst, I try to keep the focus on clarity, reliability and long-term impact rather than quick wins only.