About Me

As a Summer Analyst at Goldman Sachs

I started working on my assigned project, Global Trip Records (GTR) from scratch and completed it within 7 weeks' deadline which was challenging being a complete novice in Java and Springboot. I transitioned the ingestion process from consuming XML data from SFX, 3 times a day to fetching JSON from REST API, every 2 hours. Persisted data to DB2 & Data Lake. Alleviated challenges like High Operational Cost, Tech Debt and Resource Locking. Achieved improved continuity & frequency. Achieved >90% JUnit coverage on modular code written from scratch & designed ER schema for 600 attributes following ACID properties through extensive testing. Analysed >50k Trips & worked with Business Team and Vendor and brought to light several Data Quality & Quantity issues.

As a Founding Research Member at Unicode Research,

I am working on projects involving Probabilistic Modelling, Social Network Analysis and Causal Inference. Identified narratives spread during the span of Ukraine-Russia conflict and preformed hierarchical clustering based on it. Analysed spread of misinformation using a new metric, impact score, which is dependent on account's mean toxicity in tweets (fetched for each tweet using Perspective API), no of likes, retweets, replies received for a tweet, follower’s centrality and impact scores, etc for a particular narrative. Analysed Retweet history of grouped users for articles of interest, to identify which group of users are actively contributing to spread of misinformation than the rest. Classified accounts into 3 levels of risk representing contribution in share of misinformation articles. Among the final projects at UMLSC 2021, my Twitter Injury Report project bagged 1st place.