Inspiration
Despite their excellent skills, foreign workers face many inconveniences and restrictions due to unfamiliar language.
We were inspired by multilingual speakers using a mix of languages in one sentence. And we also focused on the fact that they are ready to continue learning and using a foreign language.
DengL helps foreign workers communicate in a mixture of their mother tongue and foreign language. So that they can take advantage of the words and phrases they have learned.
What it does
The app allows a streamlined and simplistic approach to tracking field data as a labourer. Mainly to be used on Smartphones, we provide quick voice input to take memos. For immigrants, we encourage the use of the already learned chunks of the new language by allowing textual input to be in multiple languages, all to be translated into a target language.
Furthermore, reports can be automatically summarised once they reach an adequate length. All entries can be consolidated into a PDF output.
How we built it
DengL is realised as a web-app that communicates with AI backend services via REST.
For the web frontend, we used Svelte with Vite as our build toolset and leverage GitHub pages for deployment.
Our AI driven backend uses azure, openai and cohre to master the complex task of translation and transcription. We use FastApi for our endpoint and pydantic for type validation.
Challenges we ran into
Microsoft Azure has a lot to offer, but tricky to get into Finalizing features gets more tricky the more you test it
Accomplishments that we're proud of
- aruuning mvp in the web
- great team work
What we learned
Brainstorm first, code second Stay focused and within scope Don't order 150 pizzas at once
What's next for DengL
Refining our language models Adding more features (image, video logging) Creating a business plan
Built With
- css
- fastapi
- html
- javascript
- pydantic
- python
- svelte
- typescript
- uvicorn
Log in or sign up for Devpost to join the conversation.