About
Shubham Gandhi
Data Scientist working on recommendations, ranking, and applied ML systems. This site collects technical essays on AI, ML, mathematics, and the systems thinking behind real-world products.
What I Write About
I am most interested in the gap between clean textbook ideas and messy production behavior: how ranking metrics fail, how model assumptions leak into product decisions, and how data systems shape what models can actually learn. The writing here is closer to practitioner analysis than step-by-step tutorials.
What I Build
I like small tools that reduce repeated work. One example is Jobs Digest, a packaged workflow for turning job-alert emails into a personalized jobs dashboard. The package is designed so other candidates can adapt it without inheriting my private keys or Gmail credentials: configuration is separated from secrets, and a manifest lays out the steps needed to connect their own inbox notifications, filters, and dashboard preferences.
Elsewhere
Outside the core ML work, I am curious about coffee and automating routine tasks: the small scripts, dashboards, and workflows that make daily work less repetitive and more inspectable.