SpeechForGood

I developed SpeechForGood, a gamified platform designed to make Automatic Speech Recognition (ASR) more inclusive for people with speech impediments. Our research identified that users were often discouraged from contributing speech data due to privacy concerns, a lack of transparency, and difficulty recording high-quality samples. To address these challenges, we created an interactive, web-based experience where users record voice samples to "feed" and customize a digital creature, effectively lowering the barrier to entry while building a supportive, community-driven dataset.

Throughout the project, I contributed to every phase of the design process, ranging from foundational research to high-fidelity prototyping. I conducted user interviews to uncover the specific needs of individuals with speech impediments and performed a comprehensive literature review to guide our ethical and technical standards. I then translated these insights into wireframes and a final prototype, focusing on design techniques that prioritize user agency, such as intuitive playback, re-recording capabilities, and transparent, optional data management systems.