Slivergenpal | Devpost

logo
Frontend
Fast API Endpoints
Speech-to-Text to streamline registration process

Inspiration

From my past volunteering experience at an AAC - realize that actually the ratio of male to female seniors participating in a social event is like 10-90% and after hearing from a male senior retiree himself - he mentioned how sometimes it's the nature of the activity (not interesting enough/does not involve building up a skill) and most of it is one off. (difficult to stay in sync with friends)

What it does

For seniors - platform to check if the senior's regular kakis(buddies) are showing interest for the opportunities curated to build a new skill for the male seniors by working with various organisations. Also for repeated sign-ups/filling in of forms - we have an auto-fill up system to make the process more efficient.

For the AACs - providing a standardised way to track engagement history and equipping staff the tools to easily visualise this data - easier for follow-up plannings.

How we built it:

Frontend: Next Typescript Database: Supabase (Postgresql) Backend: FastAPI, Python, Whisper (speech-to-text), Gemini API, Supabase Auth and REST service.

Deployment: Vercel

Challenges we ran into

Limited time for developing the website, backend and database services together.
Ensuring that our website is secure and our data integrity is protected by enabling row level security and the data policies to prevent other unauthorized users from accessing the data.
Need to support accurate multilingual transcription because of ethnic diversity among users while keeping the transcription latency low.
Natural language processing to map the transcribed content to the form fields, to support form autofill using our voice feature.
Figuring out the entire flow when it came to how we make sure both the AAC & our male seniors will find it helps solve their pain-points.

Accomplishments that we're proud of

Being able to create a minimum viable product that integrates both the website and the backend services together to create a fully functional flow for our users.
Also integrating software & hardware.

What we learned

Learning to empathize and building products for social good, in this case the ageing population.

What's next for Slivergenpal:

We'll continue fine-tuning the solution and partnering with active-ageing centres to potentially do a validation of our prototype and doing pilot testing.