Katalume documentation

Practice machine learning into mastery.

Katalume is the training ground for machine learning — solve real ML problems in an in-browser judge, compete in contests, and climb to mastery. LeetCode rigor meets Kaggle depth.

The name combines kata, deliberate practice that forges mastery, with lume, light or illumination—the moment a hard problem clicks.

This repository documents the current product, the frontend and backend architecture, every API, environment configuration, development workflow, production operations, security expectations, and the July 20, 2026 launch plan.

Current public beta: https://katalume.vercel.app with 198 practice problems and pinned in-browser Python execution. Server-side ranked execution remains intentionally disabled.

Quick start

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
mkdocs serve

Open http://127.0.0.1:8000.

Source of truth

• Product behavior: public frontend repository
• API behavior: public backend-api repository
• Practice content: public ml-problems repository
• Product and operational documentation: this repository
• Launch gates: docs/launch/readiness.md
• Deadline plan: docs/launch/july-20-plan.md

Documentation changes must stay aligned with deployed behavior. When an API, environment variable, workflow, or user-facing feature changes, update the corresponding page in the same delivery cycle.

Validation

mkdocs build --strict

Strict builds run in GitHub Actions on every push to main and on pull requests.