Tryit, AI virtual try-on, and graduating with First Class

In October 2024 I defended my final-year project for the Advanced Diploma in Information Technology at RP Tumba College, with my project partner ISHIMWE Lea.

The project was Tryit.

What Tryit does

Tryit is an AI-powered virtual try-on system for clothing. You capture an image of yourself, choose a piece of clothing, and the system shows you wearing it.

For online clothing sellers, the problem is simple: customers can't try clothes before they buy, so they hesitate, or buy and return. Seeing a garment on your own image reduces that uncertainty.

How it works

No single model could do the job, so we combined several. Stable Diffusion handled image generation, DensePose mapped the person's body in the photo, and other models filled in the rest of the pipeline.

The critical part was alignment: fitting the garment onto the person's image while aiming to preserve its real measurements. A try-on that makes every garment look like it fits perfectly is useless to a shopper. The clothing has to keep its true shape and size on the body.

Supervision

We built Tryit under the supervision of NIYONSHUTI Yves, one of the lecturers at RP Tumba College experienced in AI. His guidance shaped both this project and the Ndengera chatbot later that year.

Graduation

On 29 May 2025 I graduated with an Advanced Diploma in Information Technology, First Class.

It came after three years in which I had also registered HDEV TECH Ltd, won first place with Uruziga, started on messaging infrastructure, joined ITEC Ltd as a software developer and competed in several hackathons.

In September 2025 I returned to RP Tumba College to pursue a BTech in Information Technology.

What I took from it

Real problems need combined tools. Tryit worked because we chained models together, each doing the part it does best.

Accuracy is the product. A try-on that ignores measurements looks impressive and helps nobody. Aiming to preserve real sizes is what separates a useful idea from a demo.

Work and study can reinforce each other. What I built at work improved my studies, and what I learned in class improved my work.