Dating App recommending people based on their Psychology Profiles

At the end of 2022, I managed a team of 3 developers to build a recommendation system to match people on a dating app:

AWS Architecture of a Recommendation System at Scale

The recommendation engine was powered by a match-prerank-rank-rerank algorithm running on OpenSearch that I conceptualized to save computing Power, similar to the one used by TaoBao. Given that the dataset was supposed to host over 1 Million sample, I created multiple embeddings of the user data. Then, I employed a ANN (Approximate Nearest Neighbor) on a first embedding to filter the initial search, using an accurate KNN filter for the remaining embeddings.

personalized match-prerank-rank-rerank system

The Engine was built on top of OpenSearch AWS service. The rest of the project was dedicated into building the actual app, storing user data with AWS Cognito and their preferences into a DynamoDB database.

Coolest projects

Here I list some of the biggest Project I have managed:

Fractional CTO Resume