The ScaDaMaLe Book — 2026 edition (current)

A single, self-contained text on the mathematics and the engineering of scalable, distributed, LLM-era machine learning, taught together as one craft: scalable algorithms priced by their cost (work, depth, communication, memory), distributed machine learning, and large language models introduced from day one as the course’s running workload. Every core result is machine-checked, and every lab runs in a reproducible container on open components.

The 2026 book is a working draft, openly licensed CC BY-SA 4.0, and free to read and use for self-study. This is the current edition; earlier versions are listed below.

Skills map & prerequisite atlas

Which 2026 industry skill-gaps this book closes, and the minimal path through it to close each — interactive. Click a skill area to light up its pathway through the prerequisite atlas; click a chapter to open it in the free PDF.

Interactive skills-map & prerequisite atlas. Enable JavaScript to explore it — click a 2026 skill-demand area to light up the minimal path through the book that closes it, and click a chapter to open it in the free PDF. Without JavaScript:

Earlier versions

ScaDaMaLe continues an open scalable-data-science course lineage taught and refined across several editions since 2017.

  • Scalable Data Science and Distributed Machine Learning (ScaDaMaLe), 2021 — SDS-3.x
  • Scalable Data Engineering Science with Apache Spark 2.x, 2019 — SDS-2.x
  • Introduction to Data Science, Summer 2019 — in/2019
  • 360-in-525 minutes course set in the data sciences, 2018–2019 — 360-in-525

See the full list on the Courses page.

Licence

The ScaDaMaLe Book (prose, mathematics, figures, and student exercises) is © 2026 VakeWorks AB, licensed under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0) — free to share and adapt, including commercially, with attribution to VakeWorks AB and under the same licence. Runnable code distributed with the book is separately licensed (AGPL-3.0 or permissive, as marked). Instructor solutions and customised workshop materials are not distributed.

Course pathways (2021 archive)

The course was delivered as a WASP PhD course in 2020–2022 with sponsorship from Combient Mix, Wallenberg AI Autonomous Systems and Software Program (WASP), the Department of Mathematics and Centre for Interdisciplinary Mathematics at Uppsala University, and Databricks University Alliance + AWS credits. The interactive Course Pathways app below is a sovereignty-respecting local mirror of lamastex.github.io/ScaDaMaLe (US-hosted GitHub) — hosted here on VakeWorks’ EU/Hetzner infrastructure for jurisdiction-explicit access continuity.

The Course Pathways app is also available standalone at /courses/ScaDaMaLe-app/ if your browser blocks iframes.