Artificial Intelligence
Applied AI engineering — from model intuition to shipping intelligent features with real evaluation loops.
- Mode
- Online
- Level
- Intermediate
- Certificate
- Verifiable
- Stack
- 7+ tools
Hiring teams don't pay premiums for people who've watched tutorials — they pay for people who can ship reliable, evaluated AI features without babysitting. This program is your proof of that, backed by projects an interviewer can actually open and inspect.
What this program actually does.
This track exists for one reason: most people who claim to "know AI" have never shipped anything that had to work reliably in front of a real user. You'll build the intuition first — how models actually behave, where they break, why a demo that looks magical can fail quietly in production — and then spend the rest of the program turning that intuition into working software.
By the end, you won't just be able to call an API and get a response. You'll be able to design a system around a language model: retrieval, evaluation, guardrails, cost control and monitoring, the parts that separate a weekend project from something a company can actually depend on.
Who should join
- Final-year students exploring AI roles
- Developers moving into applied AI
- Analysts adding modelling depth
- Product engineers who need to ship AI features responsibly
Roles this prepares you for
Structured phases, built to compound. Here's exactly what you'll learn.
AI foundations
Vectors, probability intuition, model behaviour and failure modes.
- Build intuition for how neural networks represent meaning as vectors
- Understand why models hallucinate, and how to reason about confidence
- Learn the vocabulary you need to read any AI paper or product doc
Working with LLMs
Prompt design, embeddings, retrieval and structured output.
- Write prompts that are reliable across edge cases, not just the happy path
- Use embeddings and vector search to ground answers in real data
- Force structured, parseable output your application can trust
Agentic systems
Tool use, function calling, multi-step reasoning and orchestration.
- Design agents that call tools and APIs to complete multi-step tasks
- Handle failure, retries and loops without the agent going off the rails
- Orchestrate multiple specialised agents around one goal
Evaluation
Datasets, benchmarks, guardrails and regression testing for AI features.
- Build a labelled evaluation set so you can measure quality, not guess at it
- Set up automated regression tests that catch silent quality drops
- Add guardrails for safety, tone and factual accuracy
Fine-tuning & adaptation
When to fine-tune vs. prompt, LoRA basics and dataset curation.
- Decide when fine-tuning is actually worth it versus better prompting
- Curate and clean a training dataset that improves the model, not confuses it
- Run a lightweight fine-tune using LoRA and measure the impact honestly
Deployment
APIs, latency budgets, cost control and observability.
- Ship an AI feature behind a real API with sane latency budgets
- Track token spend and set cost guardrails before they surprise you
- Add logging and observability so you know when quality slips in production
Technology stack
Projects
- Retrieval-augmented assistant over a private document set
- Resume-to-role matching engine with evaluation harness
- Multi-step agent that completes a real task using tools
- Production API serving a fine-tuned model
Learning outcomes
- Frame a business problem as an AI problem
- Ship an evaluated, documented AI service
- Design and monitor agentic workflows responsibly
- Communicate trade-offs like an engineer, not a hobbyist
Online · Live mentorship · Intermediate
Artificial Intelligence questions.
Ready to start the Artificial Intelligence track?
Talk to our team about the track, internship or institutional program that fits where you are today.