Working developers
Backend, frontend and mobile engineers who want to build LLM applications properly, not as demos.
For developers. Build with model APIs, retrieval, agents and MCP, prove it works with evals, and deploy it with tests, tracing and cost controls.
Coding agents make bad engineering faster. This school teaches the engineering that makes AI systems correct, safe and cheap to run.

Backend, frontend and mobile engineers who want to build LLM applications properly, not as demos.
Graduates who can code and want a portfolio of deployed AI systems that employers can inspect.
People with some Python or JavaScript who want to move into AI engineering. If you are not ready yet, we will point you to Python for AI first.
Projects use African data and problems by default: mobile money, central-bank circulars, local languages, patchy connectivity.

A service that turns messy WhatsApp order messages into validated JSON, with tests and token accounting.
Retrieval over public central-bank circulars, with an eval suite of at least fifty cases and a score report.
A finance assistant that checks exchange rates, calculates budgets and drafts reports, with clear limits.
A Model Context Protocol server that exposes a business system safely, with permissions.
Receipt-to-expense parsing, or a voice FAQ in Pidgin or Swahili, with latency measured.
A brief from Boltcliq, a hiring partner or your own idea: spec, build, eval, deploy and demo.
Free projects, roadmaps and challenges in Labs.

Free projects on the AI Engineer roadmap.
The Builders track: RAG and an agent in one weekend.
AI Coding for Developers, AI Agents in a Week, AI Security Essentials.
Twelve weeks, 8 to 10 hours a week, six deployed applications.
Six months, hybrid, with a production capstone for a real organisation.
Full-time and paid, on live Boltcliq engagements.
An AI feature is not done until you can show, with a test set, that it works and keeps working.
You will use Claude Code, Cursor and similar tools. You will also plan, review and test what they write.
Prompt injection, data leakage and PII handling under African data law are part of the course, not an appendix.
AI help is allowed and expected. A short technical conversation, the viva, confirms you can explain what you built.

Start with a free project in Labs, or build your first AI solutions in the free Masterclass.
For developers. Build with model APIs, retrieval, agents and MCP, prove it works with evals, and deploy it with tests, tracing and cost controls.