📰 Medium · Machine Learning

When I first started exploring artificial intelligence, I thought it was a modern breakthrough-something born out of recent technological… Continue reading on M

The classifiers we have looked at so far usually train a model by learning a rule, formula, or decision boundary from the training data… Continue reading on Med

How a vocal separator cuts a song into six-second blocks and stitches them back without a single tick at the seams. Continue reading on Medium »

Welcome to Day 1 of my Machine Learning series! Continue reading on Medium »

Do diffusion language models sacrifice quality for speed? NVIDIA’s July 2 release cut the tax to 1.3 points on average — but code and math… Continue reading on

Cost per token is an infrastructure metric. Cost per successful task is a business metric. Continue reading on Towards AI »

The order I skipped, the traps I fell into, and the path I’d hand you if we were sitting down together. Continue reading on Medium »

For years, dashboards have been the cornerstone of business intelligence. Continue reading on Medium »

Why AI for Business Is About Empowering People, Not Replacing Them Continue reading on Medium »

Every morning I spent 40 minutes doing the same joyless thing — scanning emails, pulling news, updating a tracker, writing a brief nobody… Continue reading on A

The Assumption that breaks every first-time AI/ML Engineer Continue reading on Medium »

Everyone’s fine-tuning. Nobody agrees on how. Here’s the honest breakdown of three methods, when each one earns its complexity, and why… Continue reading on Tow

My honest, no-hype breakdown of what’s worth the money right now — and what I’m telling you to skip Continue reading on Medium »

Every speculative decoding system I’ve looked at makes the same trade. You either get a drafter that’s accurate but slow, or fast but… Continue reading on Mediu

Day 28 is a performance optimization day: profiling the entire stack (FastAPI + Redis + Postgres + ML inference), finding real bottlenecks… Continue reading on

What happens when a deep learning model trained on curated tiles meets an actual 5.5-billion-pixel whole slide image and why the gap… Continue reading on Medium

Recommender systems (also called recommendation systems or RecSys) are information-filtering systems that use data about users, items, and… Continue reading on

The industry is currently obsessed with AI agents. Continue reading on Medium »

Yesterday we split our data three ways and saw the validation set catch a network in the act of memorizing instead of learning. Today we… Continue reading on Me

Apache Kafka has become the de facto standard for real-time data streaming and event-driven architectures. But simply adopting Kafka isn’t… Continue reading on

Retrieval-Augmented Generation is not “put documents in a vector database and call an LLM.” In a regulated enterprise — PHI, PCI… Continue reading on Medium »

Three libraries. Endless possibilities. Here is how to get them installed and understand what each one actually does. Continue reading on Medium »

Multitask learning is all the rage these days, being especially useful and widely adopted in modern ranking systems like ads, search, and… Continue reading on M

How LLMs Are Quietly Transforming Everyday Business Workflows Continue reading on Medium »