The problem Stories solves

The importance of mental health has been in its rightful spotlight for a while now but many people don't have access to the costly form of therapy. It is well known that people are likely to share their experiences with a stranger rather than an acquaintance due to the fear of being judged. It is a common practice to seek help via online chatrooms. The issue is that these chatrooms are not monitored properly. The rare ones that do, can be easily used against their purpose and push the seeker in a worse place.
This is where our product stories come in. We provide a simplistic chat room that does much more for people who want to provide help and get helped.

Challenges we ran into

Implementing a chat room management system along with a banning middleware in Redis was rather difficult. Moreover, implementing authentication with JWT in Socket.io events was a second challenge as far as the web application goes. For the Tensorflow model, we had to fine-tune the model as per our applications, which meant fiddling with the confidence threshold of the output categories to get an optimum solution. It was also pretty tricky to pair the seeker & supporter based on their tags selection. After pairing them, getting a score based 'compatibility' for the same. And pushing them into the same chatroom based on the score.