About

Taught by someone who builds this for a living.

Applied AI School is run by Houston Muzamhindo. One instructor, which means the person who designs the curriculum is the person standing in front of you.

Leading AI in banking: 7+ years
Learners taught online: 27,000+
AI champions enabled: 200+

The short version

I am Head of Applied AI in international private banking, where I have spent the last several years building and scaling an AI capability from the ground up: strategy, delivery, governance and adoption. The systems I build go into production and get used by people whose work depends on them being right.

I have also been teaching for over a decade, first through IQmates and then on Udemy. Applied AI School is where those two halves meet: production experience taught properly, rather than theory taught by someone who has never shipped, or tips taught by someone who has never had to explain a model to a risk committee.

What I build

A credit-review platform built on Azure OpenAI that produces deal minutes in under ten minutes. That work previously took around two months per deal.
Agent-based systems on Azure AI Foundry that surface and qualify opportunities, identifying over £1bn for front-line teams to act on.
Production models for life-stage targeting and next-best-offer, running against real customers rather than notebooks.

Beyond building

Worked part time for OpenAI through Turing from January to April 2023, during the training of the GPT-4 model.
Lead Champion of a global AI Champions Network, enabling 200+ champions across teams, geographies and seniority levels.
AI roadshows delivered in London, Mumbai and the Channel Islands, surfacing 200+ internal use cases.
Member of an AI Fabric Council setting principles and technical build standards, with sign-off authority on AI governance.
Member of the UK Finance AI Policy Committee, contributing to AI policy for the UK financial services sector.

What I have taught

Udemy instructor since 2020, teaching Supervised Learning From First Principles and Unsupervised Machine Learning From First Principles, with 10,000+ enrolled learners and Deep Learning From First Principles in production.
Founded IQmates, an online learning platform that served 17,500+ students across 25+ courses.
Regular internal training, weekly AI drop-in clinics, and mentoring data scientists and interns from problem framing through to deployment.

How I teach

From first principles

It is in the title of nearly everything I have made (the Udemy courses, the book) because it is the whole point. Tools change every few months. The mathematics, the mechanics and the failure modes underneath them do not. If you only learn the interface, you are starting again every time the interface moves.
Taught by someone still building

I teach what I am currently doing, not what I read about. That means the examples come with the awkward parts left in: the thing that was slow, the model that was confidently wrong, the design that had to be explained to people who were not going to accept "the model said so".
You should be able to explain your own work

Use AI to help you learn. I do. But the bar is that you can account for what you built and why. Anyone can produce something that runs. Being able to say why it works, where it breaks and what it costs is the difference between a demo and a system.
Depth should not be a luxury item

Serious technical education is usually priced as though only employers should be able to buy it. Cohorts here are priced so that the deciding factor is whether you want to do the work, not whether you can afford to find out.

The book

Practical LLMOps Deep Dive

Building Production LLM Systems from First Principles

Most LLM tutorials stop at model.generate(). Most production tutorials start at vllm serve. Very few explain what happens in between, and that gap is where the real operational work lives.

A first-principles guide to deploying and operating large language models, anchored on Qwen3-0.6B with explicit scaling commentary for 7B and 70B models. Self-published, first edition.

Get the book on Gumroad

Background

MSc Artificial Intelligence (in progress), IU International University of Applied Sciences, Germany
MBA (in progress), University of Suffolk, UK
Postgraduate Diploma in Business Management, University of the Witwatersrand
BSc Computational and Applied Mathematics, University of the Witwatersrand

Why this school exists

Most AI education stops at the surface: how to use this week's tools. That is useful, and it is not enough. The people who will do well over the next decade are the ones who understand what is happening underneath, and can build things that survive contact with real users.

This school started as an idea for Zimbabwe. Interest came from too many other countries for that to make sense, so it became Applied AI School: same intent, no borders. The pathway runs from using AI, to building with it, to engineering the systems behind it.

Applied AI School is an independent venture. Courses and views are my own and are not affiliated with, or endorsed by, my employer.

Start where you are.

Create an account with Google in a few seconds. No password to remember.

Create your account Browse courses first