Loading
Loading
Classical ML, deep learning with PyTorch, NLP, LLMs, RAG, and the operational discipline to run ML systems in production.
Engineers and analysts who want a serious, end-to-end grounding in modern ML and AI — including the LLM era. We cover both the maths and the production engineering.
Three capstones: a classical ML classifier with proper evaluation, a PyTorch deep-learning project on real data, and a production RAG application with LLMs.
Live cohorts with both maths-side and engineering-side guidance. Capstones graded by senior engineers; standout projects are presented to studio clients.
You should already be comfortable with:
10 modules · 38 lessons
NumPy and Pandas — the bedrock you cannot skip.
NumPy arrays and broadcasting
Pandas dataframes for real data work
Data cleaning at scale
Jupyter notebooks done well
See your data before you model it.
Matplotlib essentials
Seaborn for statistical plots
Sign up to lock your seat. Pay in full or set up an instalment plan at checkout.
Plotting for stakeholders, not just yourself
The intuitions behind every model.
Supervised vs unsupervised vs reinforcement
Train/test/val splits and leakage
Loss functions and what they really mean
The bias-variance tradeoff in practice
The models you should reach for first.
Linear and logistic regression
Decision trees, random forests, gradient boosting
Pipelines and proper cross-validation
Evaluation metrics that match the business
Neural networks from first principles.
Tensors, autograd, and the training loop
Building MLPs and CNNs
Working with GPUs and mixed precision
Saving, loading, and deploying models
How machines read language.
Tokenisation and embeddings
Recurrent and attention-based architectures
Sentiment, classification, and entity extraction
Use modern large language models well.
The transformer architecture in plain English
Prompt engineering and structured output
Fine-tuning vs in-context learning
Safety, evaluation, and red-teaming
Build LLM apps grounded in your own data.
Embedding pipelines
Vector databases (pgvector and friends)
Retrieval strategies and reranking
Evaluation of RAG systems
Ship models the way a senior team does.
Serving models with FastAPI
Monitoring drift and model decay
Re-training pipelines
Cost, latency, and the production reality
Three projects — classical, deep, and LLM-powered.
Classical ML capstone
Deep-learning capstone with PyTorch
RAG product capstone
Present to studio engineers