Pre-booking now

AI Assistants for Coding

from first prompt to a production-grade PoC

8 weeks · Sat & Sun 19:00–21:00 (UTC+1) Live on Google Meet

AI coding assistants have changed how software gets built. Instead of writing every line yourself, you direct an AI pair programmer that writes code, finds bugs, generates tests and helps you move from idea to working system in a fraction of the usual time.

But fast is not the same as good. This course teaches you to use AI assistants the way professionals do: with version control, written specs, automated tests, security checks and a proper deployment, so what you build is something you can trust and show to others.

You start simple, with the tools inside VS Code and PyCharm, Claude, ChatGPT and open-source assistants you can run privately on your own machine. Then you put everything together in a four-week capstone: a database chatbot that lets people ask questions about their data in plain language, built with AI, tested, secured and deployed live on a VPS. Every session is hands-on, and you finish with a working project and a repository you can present to employers or clients.

Pre-book a seat View the flier
8Weeks
18Live lessons
54hHands-on hours
LiveSat & Sun online
GroupLive instruction
100%Project-based
Who it's for

Who this course is for

  • Developers and students who know basic Python and want to build faster and better with AI
  • Freelancers and small teams who want to deliver working prototypes to clients quickly
  • Anyone who has tried AI coding tools and wants a reliable, repeatable workflow instead of trial and error
Outcomes

What you'll learn

By the end, you'll have a live database chatbot running on your own server, and the skills to build the next one faster.

Master your AI pair programmer
  • Get more from Copilot, Claude, ChatGPT and JetBrains AI inside VS Code and PyCharm
  • Debug fast by giving the AI the right logs, errors and context
  • Refactor and generate tests without breaking what already works
  • Keep the AI consistent across a whole project with instruction files
Turn ideas into specs, and specs into working code
  • Write a clear spec so the AI builds the right thing the first time
  • Have the AI critique your design before a single line is written
  • Break big features into small, checkable steps instead of one giant prompt
  • Choose the right model (Claude, GPT or open-source) for each task
Delegate to AI agents, and stay in control
  • Hand off multi-file tasks and review the changes before accepting them
  • Spot when an agent has gone off track, and steer it back
  • Automate repetitive work so you spend your time on decisions, not typing
Ship with a safety net
  • Use Git so any bad AI change can be undone in seconds
  • Add automated tests and quality checks that run on every push
  • Set up a GitHub Actions pipeline so broken code never reaches your server
Lock it down before it goes live
  • Keep passwords and API keys out of your code and your history
  • Catch insecure defaults and made-up or unsafe packages that AI can suggest
  • Understand prompt injection, and why it matters when AI reads files and data
  • Give your app the least access it needs, and nothing more
Keep sensitive code and data private
  • Run AI assistants on your own machine with Ollama and Continue
  • Compare local and cloud models on quality, speed, cost and privacy
  • Classify your data, so you know what can go to the cloud and what never leaves your laptop
Let people talk to their data
  • Turn plain-language questions into safe database queries
  • Block dangerous queries, cap runtime and limit results
  • Handle vague, unanswerable and malicious questions gracefully
  • Build a test set of questions with known answers so quality never quietly slips
Build real software, not demos
  • Add login, per-user permissions and an audit trail of every question asked
  • Manage database changes safely with versioned migrations
  • Handle errors without crashing or exposing internals to users
  • Track what every AI question costs and set spending limits
  • Write documentation someone else can actually follow
Go live, and stay live
  • Package your app with Docker so it runs the same everywhere
  • Deploy to a VPS with HTTPS, a firewall and a health check
  • Set up automatic backups, test a restore, and roll back a bad deploy
  • Get an alert the moment your app goes down
  • Run a 16-point production-readiness checklist before you call it done
Build a working PoC in hours
  • Take the same workflow to a database you have never seen, against the clock
  • Show an organization a working system, not a slide deck
  • Know exactly when a proof of concept is ready to trust, and when it isn't
Syllabus

Week by week, building real skills

Two live lessons each week — a Saturday lesson introducing fresh material and a Sunday guided lab to embed it.

Week 1 · AI assistants and IDE integration

How AI coding assistants work, and what each tool is forGitHub Copilot: setup, inline suggestions and chatContinue extension: local and cloud models in VS CodeGit as your safety net: commits, branches and reverting bad changesProject setup: environments, dependencies and keeping secrets safeYou build: a working Python API in a clean Git repository

Week 2 · Claude, ChatGPT and JetBrains AI

Claude's strengths: long context, reasoning and code generationCurrent OpenAI and Anthropic models for code generation and debuggingPyCharm AI Assistant and JuniePrompting strategies for complex code tasksSpec-first development: requirements, plan, then small implementation stepsDebugging with AI, and generating and running testsYou build: a tested API with a database, created from a written spec

Week 3 · Open-source and local assistants, plus security

Ollama: running models locally on your machineContinue with local models (Llama, Mistral, Qwen)Cody by Sourcegraph: context-aware code assistancePrivacy and speed trade-offs of local vs cloudSecurity for AI-generated code: secrets, insecure defaults, unsafe packages and prompt injectionWhat never leaves your laptop: deciding which data can go to cloud models and which must stay localYou build: a local-vs-cloud comparison, a completed security review of your own code and a data classification sheet

Week 4 · Agentic workflows and automation

Agent mode: delegating multi-file tasks and reviewing the changesProject instruction files that keep AI consistent across a projectConnecting AI to tools and automating repeatable tasksQuality gates: linting, formatting and automated testsYou build: a repository with a passing GitHub Actions pipeline

Week 5 · Capstone: spec, architecture and data layer

Writing the spec with AI and having the AI critique the designAPI skeleton, PostgreSQL and Docker ComposeA read-only database role from day oneDatabase migrations, so your schema can change safely as the app growsYou build: a containerized skeleton with migrations, running on a realistic sample database

Week 6 · Capstone: natural language to SQL

Turning plain-language questions into SQL with an LLMQuery validation, timeouts and row limitsHandling ambiguous and unsafe questionsAn automated evaluation set to catch regressionsYou build: a chatbot engine that passes its own test suite

Week 7 · Capstone: interface, users and observability

A web chat interface with tables and simple chartsLogin and per-user permissionsAudit logging, error handling and configurationTracking and capping AI costs: know what every question costs and set spending limitsYou build: a complete application with authentication, a full audit trail and cost controls

Week 8 · Capstone: deploy, harden, document and demo

Deploying to a VPS with Docker, HTTPS and a firewallBackups and rollbacks: automated database backups, a tested restore and a way to undo a bad deployUptime monitoring with an alert when your app goes downHardening, documentation and a final production-readiness reviewThe timed challenge: build a working PoC against a database you have never seenDemo dayYou build: a live, documented, deployed system with backups, monitoring and a recorded demo
Capstone

Production-ready checklist

Your capstone project is graded against this 16-point checklist.

  • Code is in Git with meaningful commits and no secrets in history
  • Database access uses a read-only role with least privilege
  • Every generated SQL statement is validated before execution
  • Query timeouts and row limits are enforced
  • Authentication and per-user permissions work
  • Every question and query is written to an audit log
  • An automated evaluation set and unit tests pass in CI
  • Linting and formatting checks pass in CI
  • The application runs from Docker Compose with one command
  • It is deployed on a VPS over HTTPS with a working health check
  • Errors are handled gracefully and logged, with no stack traces shown to users
  • A README lets someone else install, run and understand it
  • Schema changes are managed through versioned migrations
  • Database backups run automatically and a restore has been tested
  • AI usage and cost per question are logged, with a spending cap
  • Uptime monitoring sends an alert when the app fails
Live schedule

Your live classroom

Times are shown in UTC+1 (the academy's default). Add your phone number to your account to see every lesson in your own local time. Lessons run Saturdays & Sundays, 19:00–21:00 UTC+1.

Free demo & orientation

Wednesday 30 September 2026 19:00–21:00 3 days before the course starts

Join us before anything begins — meet your instructor, watch a live sample lesson and get all your questions answered. No commitment required.

Join the demo
Week 1

AI assistants and IDE integration

Sat3 Saturday 03 October 2026 19:00–21:00UTC+1
Sun4 Sunday 04 October 2026 19:00–21:00UTC+1
Week 2

Claude, ChatGPT and JetBrains AI

Sat10 Saturday 10 October 2026 19:00–21:00UTC+1
Sun11 Sunday 11 October 2026 19:00–21:00UTC+1
Week 3

Open-source and local assistants, plus security

Sat17 Saturday 17 October 2026 19:00–21:00UTC+1
Sun18 Sunday 18 October 2026 19:00–21:00UTC+1
Week 4

Agentic workflows and automation

Sat24 Saturday 24 October 2026 19:00–21:00UTC+1
Sun25 Sunday 25 October 2026 19:00–21:00UTC+1
Week 5

Capstone: spec, architecture and data layer

Sat31 Saturday 31 October 2026 19:00–21:00UTC+1
Sun1 Sunday 01 November 2026 19:00–21:00UTC+1
Week 6

Capstone: natural language to SQL

Sat7 Saturday 07 November 2026 19:00–21:00UTC+1
Sun8 Sunday 08 November 2026 19:00–21:00UTC+1
Week 7

Capstone: interface, users and observability

Sat14 Saturday 14 November 2026 19:00–21:00UTC+1
Sun15 Sunday 15 November 2026 19:00–21:00UTC+1
Week 8

Capstone: deploy, harden, document and demo

Sat21 Saturday 21 November 2026 19:00–21:00UTC+1
Sun22 Sunday 22 November 2026 19:00–21:00UTC+1
Week 9

Live lessons

Sat28 Saturday 28 November 2026 19:00–21:00UTC+1
Sun29 Sunday 29 November 2026 19:00–21:00UTC+1
Classroom & community

Learn together, online

Live Google Meet classroom

Every lesson happens in a private Google Meet room with screen sharing, code-alongs and instant help.

The room link is shared with booked students before the course starts.
WhatsApp study group

Stay connected between sessions — homework help, classmates and weekly reminders in one group.

Get the group invite when you'd like to join.
Toolkit

What you'll work with

VS CodePyCharmClaudeChatGPTGitHub CopilotContinueCody (Sourcegraph)OllamaGit and GitHubGitHub ActionsDockerPostgreSQLFastAPI

Ready to start building?

Pre-book your seat for the 3 October cohort — or join the free demo and see for yourself.

View the flier