AI-POWERED SPEECH THERAPY

FUNWATTLE

TYPE Mobile App Dev

DURATION 3 MONTHS [JUN - SEP 2025]

ROLE Software Dev & Product Owner

SKILLS HONED AI Pipelines, Product Ownership, Frontend and Backend Dev

SKILLS GAINED Ethical AI Design, Agile Workflow, Vector Search & RAG

OVERVIEW

The project involved developing an AI-powered speech therapy app for children in collaboration with FunWattle, an early-stage startup, as part of COMP3900 (Computer Science Project). The goal was to improve accessibility to speech therapy through interactive, personalised support outside clinical settings.

As Product Owner in the Scrum team, I:

Defined and prioritised the product backlog
Translated client needs into technical requirements
Led sprint planning and aligned design and development

As a Software Engineer on the team, I:

Led the development of the AI speech therapist pipeline, including developing the RAG, prompt engineering, using vector comparisons to determine the correctness of the user's answer
Designed and built multiple frontend components and screens, including user dashboards, module libraries, speech exercise components

We focused on child-safe, accessible and intuitive UX, personalised adaptive learning experiences, and engaging, child-friendly interfaces and gamification.

Developed a functional MVP featuring:

A virtual therapist interface guiding structured exercises
AI-driven personalised recommendations
Adaptive difficulty based on user progress
A child-friendly, intuitive UI/UX

The project delivered a functional prototype aligned with FunWattle's goals, demonstrating how AI could enhance accessibility in speech therapy while maintaining ethical safeguards. Demonstrated scalable and extensible architecture for future integration. Received strong client feedback on usability and ethical design.

Beyond technical output, this project strengthened my ability to:

Lead within an Agile team
Translate healthcare needs into technical systems
Design AI solutions responsibly in sensitive domains

PHOTO GALLERY