3 Education Tools · 6 Learning Tracks · 39 Ollama Models

Learn by building, not watching

Three tools for quizzes, code challenges, and skill tracking. Powered by the SM-2 spaced repetition algorithm and 39 local Ollama models for AI hints.

View ToolsLearning Tracks
3
Education Tools
SM-2
Spaced Repetition
39
Ollama Models
6
Learning Tracks
Tools
Three education tools
Quizzes that adapt, code challenges with AI hints, and a skill tracker that maps your progress.
Quiz Platform
SM-2 spaced repetition algorithm for adaptive difficulty. Tracks recall accuracy, adjusts intervals, optimizes review schedules. Progress persists in SQLite.
quiz_platform.py
Code Challenge
Code execution sandbox with Ollama-powered hint generation. Multi-language support. AI reviews your code and suggests improvements without giving answers.
code_challenge.py
Skill Tracker
Competency mapping with skill tree visualization. Learning path optimization based on prerequisites and completion data.
skill_tracker.py
Curriculum
Six learning tracks
Each track maps to real infrastructure. You learn by operating the systems you're studying.
Python Fundamentals
400+ shell scripts rewritten as Python tools. Quiz platform tracks mastery of stdlib, data structures, algorithms.
quiz_platform.py
Systems Programming
Fleet management scripts — power optimization, service lifecycle, cron automation across 5 Raspberry Pis.
code_challenge.py
AI/ML Pipeline
Ollama model deployment, prompt engineering, Hailo-8 inference (52 TOPS), CECE custom models.
39 models
DevOps & Infrastructure
Docker Swarm on Octavia, WireGuard mesh, Cloudflare tunnels (18 tunnels, 95+ Pages), systemd services.
skill_tracker.py
Web Development
334 web apps on Lucidia, 48+ custom domains, plain HTML/CSS brand system, Cloudflare Workers.
334 apps
Security
6 security tools — password manager, SIEM, certificates, incident response, secret scanning, identity provider.
6 tools
Progress Dashboard
Learning metrics
Review Stats
Total cards reviewed
2,847
Retention rate
91.3%
Current streak
34 days
Cards due today
18
Avg session
22 min
Skill Progress
Active learning paths
4 / 6
Modules completed
38 / 90
Code challenges solved
127
Skills unlocked
64
AI hints used
203
AI Tutor
39 models, zero API calls
The code challenge hint engine runs entirely on local Ollama models across the fleet. CodeLlama for code review, Llama 3 and Mistral for explanations. Your learning data never leaves your machines — no OpenAI, no cloud inference, no per-token billing.
16 Ollama models (4 custom CECE) · CECE API · TTS API
Hailo-8 · 26 TOPS
11 Ollama models · 1TB NVMe · Docker Swarm leader
Hailo-8 · 26 TOPS
6 Ollama models · Portainer · Headscale
CPU inference
6 Ollama models · Ollama Bridge (SSE proxy)
CPU inference
Algorithm
SM-2 Spaced Repetition
The quiz platform uses the SuperMemo SM-2 algorithm to calculate optimal review intervals. Items you struggle with appear more often. Items you know well space out to days, then weeks.
EF' = EF + (0.1 - (5 - q) * (0.08 + (5 - q) * 0.02))
where q = quality of response (0-5)
EF = easiness factor (min 1.3)
interval(1) = 1 day
interval(2) = 6 days
interval(n) = interval(n-1) * EF
Related
Go deeper
AI Agents & Ollama
Research & Math
Game Engine (ECS)
228 Databases
Hardware Fleet
Security Tools