evasion

Towards Data Science (TDS) is an online publication that publishes technical articles, tutorials and explanatory essays on data science, machine learning, deep learning, large language models, agentic AI and programming. Content is organized into topic categories such as Data Science, Machine Learning, Deep Learning, Artificial Intelligence, Large Language Models, Agentic AI, Programming and Writing, with each piece carrying an author byline, publication date and estimated reading time.

The site groups content into recurring formats: a "Latest" feed, "Editor's Picks", long-form "Deep Dives" (typically 12-21 minutes of reading time, and occasionally longer), and "The Variable" newsletter written by the TDS Editors, which highlights monthly must-reads and commentary on industry topics. Representative published pieces cover the mathematical origins of the sigmoid function, random forest and bagging behavior, retrieval-augmented generation rerankers, speculative decoding on CPUs, survival analysis and the Cox proportional hazards model, recursive SQL common table expressions, LLM watermarking and hallucination detection, differential privacy and federated learning for medical data, and practical guidance on working with AI coding agents.

Business model

The publication operates a contributor-based editorial model: outside practitioners submit drafts through a TDS Contributor Portal, with in-house TDS Editors providing formatting guidance and editorial curation such as Editor's Picks and a newsletter.

Latest developments

On 2026-08-25 the TDS Editors announced that the website and contributor portal had been completely rebuilt, described as a faster site with a new portal for readers and writers; a companion guide on formatting drafts in the new portal was published the following day. New articles continued to be published daily through late August 2026.