Key Strategies for Building Effective Machine Learning Systems
A strategic approach to selecting the right model, better features and evaluation.
A strategic approach to selecting the right model, better features and evaluation.
Accurate prediction of fuel consumption in internal combustion engine vehicles (ICEVs) from real-world telematics data.
Orthogonalization is the principle of separating concerns in our ML development process, ensuring that each “knob” we tune affects only one aspect of performance, without unintended side effects on ot
Getting the algorithm right is half the battle, knowing how to tune it, normalize it, and deploy it is what separates research code from production systems.
Gradient descent is just the starting point — the real question is how fast and how reliably you can reach a good minimum.
From foundational Deep Learning training techniques to the algorithms powering modern Agentic AI. Part 1 of 4.
A practical guide to reproducible analysis pipeline.
I had a project that required a lot of work with Google Datastore. However, there is surprisingly few instructional materials for this part. So I hope my article could make the job easier!
This is a blog post for my first data analysis project: Spotify EDA, and also my first post featured on CodeX publication.