Zana is offering its VOICE-BIOME platform to power a dedicated challenge at the Bridge2AI Voice Symposium Hackathon, enabling researchers and developers to explore AI-powered voice biomarker analysis at scale.
The Hackathon is taking place on May 6, 2026, in St. Petersburg, FL, immediately following the two-day Voice AI Symposium. The hackathon is open to both programmers and non-programmers across the Challenge Track and Education Track.
Participants can compete in a dedicated challenge by leveraging VOICE-BIOME's end-to-end analytical environment to uncover diagnostic signal within the Bridge2AI-Voice dataset and build machine learning models across respiratory, voice, neurological, and mood disorders.
This challenge invites teams to push the boundaries of what voice data can reveal about human health.
The featurized dataset has been imported into VOICE-BIOME's feature extraction and analytics environment, removing significant setup friction for teams.
What participants can do with VOICE-BIOME
- Explore acoustic and linguistic features across the various disease groups.
- Analyse feature distributions stratified by speech task, including picture description, free speech, rainbow passage, sustained phonation, and more.
- Perform comparative analyses, such as male vs. female, across disease groups or speech tasks in the adult and paediatric sets.
- Train and evaluate machine learning models for differential diagnosis using built-in classification pipelines with detailed performance metrics and reporting.
- Export results and visualizations including ROC curves, box plots, and summary statistics.
For the B2AI dataset specifically, teams can explore differential diagnosis across four disorder domains: respiratory, voice, neurological, and mood disorders. Task-level performance breakdowns surface which speech tasks carry the strongest diagnostic signal.
“Bridge2AI represents one of the most rigorous voice datasets assembled to date,” states Dr. Julia Hoxha, CEO of Zana Technologies. “Making VOICE-BIOME available for this hackathon lets the participants focus on discovery rather than data wrangling. Our aim is to accelerate what the research community can do with voice as a clinical signal.”
Beyond this initiative, VOICE-BIOME is designed as a disease-agnostic foundation platform, supporting research across cardiovascular and respiratory health, neurological and mental health, behavioral and wellness outcomes, and preventive and longitudinal monitoring. It targets academic and clinical research groups, digital health and life science companies, and innovation labs exploring voice as a biomarker.
For hackathon participants and teams interested in VOICE-BIOME, the following resources are available:
