Company website : https://www.quackquack.in/
Leading Indian dating app with 14M users. 1000+ new users per day and growing. Promises veriﬁed proﬁles of users, Interest based matchmaking, and fake & spammer free app
Classifying user proﬁle information into vague, spam, infomercial, genuine. Manual veriﬁcation of user proﬁles (5 employees, 35k/month) is error prone, tedious and boring
- Hamon developed machine learning techniques to create models that can help detect bad actors from genuine users and improve the usability of the app.
- Hamon was able to come in and quickly develop the right machines learning models and help QuackQuack
82% Accuracy rate in detecting spams based on available models and timeframe of the project. The company was able to remove human curators
SciKit Learn, Pandas, Jupyter Notebook, AWS EC2, Python webstack (Flask, PostgreSQL, SQL Alchemy)