A structured, hands-on personal curriculum through AI and Machine Learning Engineering - from machine learning foundations and neural networks to transformers, foundation models, applied AI systems, and production deployment.
This repository originally focused mainly on the applied AI layer: LLM APIs, embeddings, RAG, agents, MCP, orchestration, and deployment.
That stage was largely completed through the DataCamp Associate AI Engineer for Developers track, alongside additional hands-on work and projects.
The direction has now shifted deeper.
The goal is to build stronger foundations in:
Machine Learning
↓
Neural Networks
↓
Deep Learning
↓
Transformers
↓
Foundation Models
↓
Applied AI
↓
ML Systems
The objective is to understand not only how to build with AI systems, but also the models, mathematics, training, fine-tuning, evaluation, and infrastructure underneath them.
| Phase | Focus | Main Resources |
|---|---|---|
| 0 | Mathematics for ML | Mathematics for Machine Learning and Data Science — DeepLearning.AI |
| 1 | Classical Machine Learning, scikit-learn, model evaluation | Machine Learning Specialization — Andrew Ng |
| 2 | Neural networks, backpropagation, autograd, fundamentals from scratch | Neural Networks: Zero to Hero — Andrej Karpathy |
| 3 | Deep Learning and PyTorch | Deep Learning Specialization — DeepLearning.AI · fast.ai |
| 4 | Transformers, tokenization, embeddings, attention | Karpathy material · Hugging Face LLM Course |
| 5 | Open-source models, fine-tuning, LoRA/QLoRA, quantization, evaluation | Hugging Face LLM Course |
| 6 | Applied AI: RAG, agents, tool calling, MCP, orchestration | DataCamp Associate AI Engineer for Developers · DeepLearning.AI · Cohere LLM University · LangChain Academy |
| 7 | ML systems, serving, MLOps, deployment, monitoring | Made With ML · Full Stack Deep Learning · MLOps Zoomcamp |
| 8 | Computer Vision and Edge AI | Stanford CS231n and related resources |
For important concepts, the goal is to move through four levels:
understand → implement → use → integrate
That means understanding the theory, implementing simplified versions where useful, using production frameworks such as scikit-learn and PyTorch, and finally integrating models into real backend and AI systems.
Python · NumPy · Pandas · scikit-learn · PyTorch · Hugging Face · OpenAI · Anthropic · FastAPI · PostgreSQL · Redis · Docker · MLflow · DVC · LangChain · LangGraph
I also participated in a structured 10-Day AI/ML Engineering and Agentic Systems Sprint with NGOT Giants, with hands-on project work.
See 10-Day AI/ML and Agentic Systems Sprint.
I document concepts, implementations, experiments, and lessons throughout the repository for my own learning and for anyone who may find the material useful later.
🚧 In progress
Started: April 2026
Initial focus: Applied AI engineering Current direction: deeper ML, deep learning, foundation models, and production AI systems