GSoC 2026 | Trainable ChatBot Interface & AI Component | Toluwani Akintoye

Hello everyone,

My name is Toluwani, and I am a full-stack software engineer with a strong background in frontend architecture (JavaScript/TypeScript) and AI Agent integrations. I am writing to express my strong interest in the "Trainable ChatBot interface and AI component" project for GSoC 2026.

Recently, I have been architecting automated backend workflows and integrating AI agent behaviors into mobile and web applications. I am very familiar with the logic required to build a customizable model interface, and I am excited about the opportunity to build a seamless GUI for App Inventor users to train and import their own ChatBots.

I am currently setting up my local App Inventor development environment and looking through the issue tracker for a "good first issue" to tackle.

@Natalie Lao — As I begin outlining the architecture for my proposal, do you envision the ChatBot training UI functioning similarly to the existing Personal Image Classifier (PIC) standalone site, or are you looking for a more tightly integrated designer component?

(Note: I also went to check out the MIT application form linked in the wiki, but it currently says it is not accepting responses. Just wanted to give a quick heads-up in case the toggle needs to be flipped for this year!)

Thank you, and I look forward to contributing to the community!

Toluwani Akintoye
GitHub
Portfolio

Just a quick update! While setting up my local environment and looking through the tracker, I went ahead and submitted my first small contribution.

I opened PR #3831 to fix Issue #3729 (the A11y contrast bug on the Download APK dialog).

Looking forward to diving deeper into the AI Agent component architecture this week!