The Course
In 2024, I was invited to design and deliver a course on AI-powered business development at Corvinus University Budapest. The challenge: take students with varying technical backgrounds and help them build real AI products in one semester.
The course combined three threads:
- Technical foundations: Python, APIs, RAG systems, cloud deployment
- Business thinking: Value propositions, market analysis, pricing
- Hands-on building: Each team built and deployed a working AI product
What Worked
Real Products, Not Toy Projects
Every team built something that could actually be used. One team created a RAG-powered study assistant for their own department. Another built a customer support chatbot for a local business. The constraint of building for real users forced better design decisions than any textbook exercise.
AI Tools for Learning AI
We used AI-assisted coding tools throughout the course. Students who had never programmed before were deploying cloud applications by week 6. The meta-circularity of using AI to learn about AI wasn’t lost on anyone.
Guest Sessions
Bringing in practitioners – founders, engineers, product managers – gave students a reality check on what building AI products actually looks like outside academia.
What I’d Change
More Time on Data Quality
Students consistently underestimated how much time goes into data preparation. Next iteration of the course needs a dedicated module on data collection, cleaning, and validation.
Earlier Cloud Deployment
We deployed too late in the semester. Getting applications live early creates a feedback loop that improves learning. Next time, I’d have a minimal deployment by week 3.
The Bigger Picture
Teaching this course reinforced my belief that AI literacy isn’t just for engineers. Business students, designers, and domain experts all benefit from understanding what AI can and can’t do. The best AI products come from teams that combine technical capability with deep domain knowledge.
The future of AI education isn’t about training everyone to write Python. It’s about giving everyone enough understanding to participate meaningfully in building AI-powered solutions.