AI Generalist Certification

How modern AI actually works

You build: A model-selection cheat sheet for your own role

• Tokens, context windows and why models hallucinate
• Choosing between GPT, Claude, Gemini and open models
• Cost, speed and privacy trade-offs at work

Practice: Run the same task on two models and write down where each one failed.

Prompting foundations

You build: Five reusable prompts for tasks you do every week

• Structure: role, task, context, constraints, format
• Examples and rubrics that raise output quality
• Iterating instead of starting over

Practice: Turn your three most repeated requests into saved prompts.

Context engineering

You build: A prompt pack that works with your own documents

• Feeding documents, spreadsheets and policies into a model
• Chunking long inputs and keeping answers grounded
• When context beats a better model

Practice: Build a context pack for one recurring process you own.

Research and verification

You build: A sourced research brief on a topic from your work

• Deep-research workflows across tools
• Checking claims, dates and numbers against sources
• Recognising confident, wrong answers

Practice: Fact-check an AI-written brief and log every correction.

Writing that sounds like you

You build: Templates for your reports, proposals and updates

• Capturing voice and house style in a prompt
• Drafting, editing and shortening long documents
• Translation and Hinglish-to-English polish

Practice: Rewrite a real document with your template and compare time taken.

Data and spreadsheets

You build: An analysis of a real export from your systems

• Cleaning and reshaping messy data with AI
• Asking questions of a spreadsheet safely
• Charts, summaries and the errors to watch for

Practice: Produce one chart and one insight from your own data.

Automation foundations

You build: Your first working automation, end to end

• Triggers, actions and connections in n8n and Make
• Connecting email, sheets, chat and CRM
• Testing safely before you switch it on

Practice: Automate one task you currently do by hand every week.

Automation with AI steps

You build: A workflow that classifies, extracts and drafts

• AI steps inside automations: classify, extract, summarise
• Human-in-the-loop approval before anything is sent
• Error handling, retries and alerts

Practice: Add an approval step and a failure alert to your automation.

Building AI agents, part 1

You build: A simple agent that uses two tools

• What makes an agent different from a chatbot
• Tools, instructions and limits
• Planning and multi-step tasks

Practice: Give your agent a third tool and note where it goes wrong.

Building AI agents, part 2

You build: A knowledge assistant over your own documents

• Retrieval over company knowledge, in plain terms
• Testing an agent on real examples before trusting it
• Deciding what an agent may never do alone

Practice: Write ten test questions and score your assistant on them.

Vibe coding: shipping a small app

You build: A working internal tool, deployed and shareable

• From idea to app with AI coding tools
• Data, sign-in and deployment basics
• Knowing when to hand over to engineers

Practice: Share your app with a colleague and collect their feedback.

Capstone and demo day

You build: Your finished capstone, presented to the cohort

• Polishing your capstone
• Measuring the time or quality gained
• Presenting your work and next steps