Piloting FundaAI: Turning African Curiosity into Engineering Talent

I run an AI School; FundaAI to build young, self-driven and technically minded Africans between 12 and 18, to a pipeline of engineers to outsource to US companies.

Why?

• When I moved to the US, I founded an EdTech; The Gradient Boost, that raised $50,000 to train African tech talent, an experience that deepened my conviction to build tools to help convert self-learning curiosity to “cracked” engineering talent

• I have had the experience of working with and interviewing remote contractors, helping me build a better understanding of what companies tend to look for when outsourcing and the challenges they encounter

What is FundaAI

FundaAI is a two pronged AI School.

1. The virtual campus is a core component. This comes in the form of a laptop with an offline AI brain connected to an ecosystem of AI Tutors, helping kids build problem solving and fundamental software engineering skills. Why do kids need a laptop for this? Building “cracked” engineers is about instilling habits. We have a plethora of free educational content and yet, we aren’t producing more “educated people”. MOOCs create the illusion of learning; real progress happens when students apply knowledge to real-world problems. Africa doesn’t lack talent, it lacks a system to turn curiosity into engineering discipline. The laptop, loaded with AI Tutors and stripped of distractions, becomes a focused hub where kids build habits, projects, and discipline.

2. A virtual apprenticeship. Beyond the virtual campus, kids are mentored by experienced engineers from around the world. Kids work on existing code bases (ie. the FundaAI code base), getting tasks and code review to simulate a real working environment. Learning doesn’t happen in a classroom, it happens when you apply what you learn (from the virtual campus) to real world problems. With your focus corrected and directed by more experienced engineers.

Current Progress

For the past two months, I’ve been piloting FundaAI with my 12-year-old nephew, Bonginkosi, in Zimbabwe.

He’s a smart kid with an ambition to become an engineer. The MVP for the virtual campus runs on this architecture:

• Linux Mint operating system for customization and adding guard rails on what content he can view online

• The FundaAI App Store, to distribute the AI Tutors, send updates when necessary and manage user subscriptions

• The Examiner; an AI Tutor that helps kids improve their grades, through practicing past papers. The AI Tutor marks past paper questions giving students feedback on areas of improvement. The rationale behind this is simple, while kids are learning how to become engineers, it is imperative that they get good grades.

• The Engineer; The newest addition to FundaAI. This is an AI Tutor to help kids build engineering projects guided by AI. It returns projects tailored to their context and interest, breaking them down into “tickets” with AI prompts they can pass into their IDE to “vibe code” and learn while building

• Cursor; The laptops are preloaded with Cursor to enable kids to start building software guided by AI

The focus is on offline first tools to enable usage throughout power outages and factoring in high internet costs in many parts of Africa. For example, Zimbabwe has the highest data costs in the world - this is currently where I am running my pilot with my nephew. For the technically inclined, these are a combination of PySide6 and Electron desktop apps. PySide6 to leverage the visualization libraries offered by Python (highly necessary to for exam questions) and electron for the wider ecosystem and support.

Running the pilot

Part of the pilot has taken the form of weekly calls (the virtual apprenticeship) where I;

• give him tasks to work on

• I walk him through the process of debugging an AI Tutor based on a bug he identifies

I have used these an opportunity to both understand how to build the a virtual apprenticeship that actually works and figure out what to build to support the learning process. Through these sessions, I realized that we needed an AI Tutor to get kids building software on their own guided by AI; The Engineer. The mentorship sessions become more valuable when users are constantly building software and putting into practice things they learn from the “apprenticeship”.

Would love to chat to engineers who want to contribute to the code base or in helping mentor users. The end-goal is to find 1-2 co-founders to help support with building up FundaAI and building partnerships with companies for kids to shadow real world engineers on small tasks. Feel free to reach out to me at emmanuel@emmanuelsibanda.com, connecting on LinkedIn or X