A 25-session course on language and vision systems combined. NLP, computer vision, and the multimodal models that power Claude, Gemini, and GPT-4o.
From tokenisation through attention — the mechanics, not just the concept.
On a custom dataset, then probe it for bias and the features the model actually learned.
Using embeddings to find documents by meaning — the primitive behind every modern retrieval system.
Answering real questions from a knowledge base with measurably fewer hallucinations than the base model.
Captioning, visual Q&A, OCR — the architecture pattern behind today's frontier models.
Fewer than 3 in 100 applicants are selected.
Foundations and CS Specialist
"You can't build AI you don't understand. My job is to make the maths underneath it feel obvious - not intimidating."
Programming & Applied AI Specialist
"Students learn AI fastest when they're building something real with it. We start writing code on day one."
Programming & AI Engineering Specialist
"Real industry experience changes how you teach. I show students how AI is actually built, not just how it's described."
In the final five sessions, they build one complete multimodal AI application combining language with vision or audio — presented live with a real demo and written analysis of strengths and failure modes.
In the final five sessions, they build one complete multimodal AI application combining language with vision or audio — presented live with a real demo and written analysis of strengths and failure modes.
25 private 1-on-1 sessions with an IIT graduate mentor.
for the full 25-session course
$40 per 1-on-1 session · 25 sessions total
