Overview
Intelligence is not magic — it's mathematics, data, and code. From your first neural network to deploying production ML pipelines — every concept is taught hands-on, with real Docker-verified code and cybersecurity-themed examples.
🗺️ Choose Your Level
🌱 Foundations
AI history, how LLMs work, prompt engineering, RAG, agents, ethics, safety, open-source vs closed models. 20 conceptual labs — no code required, code samples included.
⚔️ Practitioner
Hands-on ML: regression, trees, neural networks, NLP, embeddings, fine-tuning, RAG chatbots, AI agents, anomaly detection, multimodal AI, and deploying with FastAPI. All 20 labs Docker-verified.
📋 Curriculum Overview
Understand AI before you build with it
01–03
History of AI, how AI works, ML taxonomy
04–06
Data and bias, neural networks demystified, transformers and attention
07–10
LLMs explained, prompt engineering, AI agents, OpenClaw platform
11–14
Vision AI, AI in the real world, AI ethics, safety and alignment
15–18
Open-source vs closed, developer toolkit, building RAG, AI in cybersecurity
19–20
AI landscape 2025–2026, capstone: design your own AI product
No code execution required — includes illustrative Python/PyTorch code samples throughout
Build and train real ML models — all Docker-verified
01–03
Linear/logistic regression, decision trees + random forests, gradient boosting + XGBoost
04–06
Feature engineering, model evaluation + metrics, neural networks from scratch (NumPy)
07–09
Convolutional neural networks, transfer learning, text classification with BERT
10–12
NER + information extraction, sentiment analysis pipeline, embeddings + semantic search
13–15
Fine-tuning with LoRA, RAG chatbot, AI agents (ReAct pattern + tool use)
16–18
Time series forecasting, anomaly detection (Isolation Forest + UEBA), multimodal AI
19–20
Deploying ML with FastAPI + Docker, end-to-end ML pipeline capstone
Docker image: zchencow/innozverse-ai:latest · Theme: cybersecurity scenarios throughout
Production-grade AI engineering
01–05
PyTorch deep dive, custom training loops, advanced CV pipelines, object detection
06–10
LLM API integration, streaming, function calling, LangChain, vector databases
11–15
Building AI apps, adversarial examples, model robustness, prompt injection defence
16–20
MLflow experiment tracking, model versioning, distributed training, production capstone
Coming soon — Advanced level labs in development
Design AI systems that scale and comply
01–05
MLOps platform architecture, model serving at scale, vector databases, LLM infrastructure, RAG at scale
06–10
AI observability & monitoring, federated learning, multi-agent systems, AI security red team, EU AI Act compliance
11–15
AI cost optimization, enterprise AI platform, data pipeline architecture, responsible AI audit, LLM fine-tuning infrastructure
16–20
Knowledge graph + LLM, real-time AI inference, AI SOC automation, distributed training, capstone enterprise AI security platform
20 Labs available — All Docker-verified with enterprise architecture patterns
⚡ Lab Format
Every Practitioner lab follows a consistent, verified format:
Each lab includes:
🎯 Objective — what you'll build and why it matters operationally
📚 Background — the theory and intuition behind each algorithm
🔬 8 step-by-step instructions — from environment setup to production capstone
📸 Verified output — real terminal output captured from Docker runs
💡 Tip callouts — explains why, not just how
🔒 Security theme — all examples use cybersecurity datasets (CVEs, SIEM, network traffic)
🐳 Quick Start
Every lab runs in a pre-built Docker image — no environment setup:
Then follow any Practitioner lab — all commands run inside this container.
All Practitioner labs are compatible with Google Colab (free tier):
Open any lab, copy the code blocks into Colab cells — they run as-is.
🏆 Certifications Aligned
AWS ML Specialty
Foundations + Practitioner
Google Professional ML Engineer
Practitioner + Advanced
Azure AI Engineer Associate
Foundations + Practitioner
Databricks ML Professional
Advanced + Architect
TensorFlow Developer Certificate
Practitioner
🔒 Cybersecurity Theme
All Practitioner and Advanced labs use cybersecurity-relevant datasets:
Network intrusion detection (classifying attack vs benign traffic)
SIEM log anomaly detection (isolation forest on security events)
CVE severity prediction and threat intelligence
Malware classification from PE file features
SOC alert triage with multimodal AI (text + screenshot analysis)
This makes concepts concrete for security professionals and adds real-world context for ML engineers wanting to enter the security space.
🚀 Start Here
New to AI? Start with Lab 01: The History of AI — no prerequisites needed.
Know Python, want to build ML models? Jump to Lab 01 Practitioner: Linear & Logistic Regression.
Security professional wanting AI skills? Start with Lab 17 Practitioner: Anomaly Detection for Security Logs.
Want to deploy a model? Go to Lab 19: Deploying ML with FastAPI + Docker.
Last updated