Datascience

A series of principles and practical habits for building, managing, and sustaining computational and data science projects with clarity and reproducibility.

Data is never clean, and that’s okay. Understanding its noise, origins, and structure tells us a lot about the data.

Data scientists should have domain-driven thinking and why understanding the problem domain still matters more than the model even in the rise of LLMs.

OpenAI’s removal of GPT-4o, o3, and other models after GPT-5’s launch breaks fundamental MLOps principles. Without model versioning and control, data science workflows become unreliable. Local LLMs offer a better alternative for maintaining consistency.

Adapting Shape Up methodology for data science work, what works and what doesn’t

I wanted to get good at Azure that I was using at work and decided to upskill and do a certification journey.