Jul 2025 – Aug 2025 · 1 mo · Lucerne

• I took a big step in my career. I left my data science job to pursue a career in neuroscience. I wanted some time to travel and reflect upon things, and a beautiful European summer is a good way of doing it. Well, at least that was the plan. I did hike in Switzerland, tasted every possible beer the Germans offer, and had a lot of fun.

Feb 2025 – Jun 2025 · 4 mos · Gurugram, Haryana, India · On-site

• Published patents with the team on the use of GNNs in the transaction world.
• Published papers with the team on the distillation of GNNs to MLPs.
• Worked on the model to detect fraudulent merchants, successfully deployed the product with improved accuracy compared to the previous version.
• Learned how to work in a large team and how to handle end-to-end pipelines in big projects.

Jun 2023 – Feb 2025 · 1 yr 8 mos · Gurugram, Haryana, India · On-site

• Worked on massive transaction Bipartite graphs to predict the risky merchant behaviours.
• Participated in hackathons to increase my knowledge and team collaboration skills.
• Collaborated in the research domain on graph ML with IIT Delhi

May 2023 – Jun 2023 · 1 mo · Himachal Pradesh

• Went on a break in the Himalayas to learn yoga, meditation and kathak dance form.

Aug 2022 – Apr 2023 · 8 mos · Mumbai, Maharashtra, India · On-site

• Worked with Prof Ganesh Ramakrishnan, Pradeep Shenoy (Google Research), and Durga Sivasubramanian (PhD, IITB)
• Performed experiments of Domain Generalization and Domain Adaptation on the WILDS Dataset
• for our research hypothesis to improve accuracy on OOD (out-of-distribution) samples in PyTorch.
• Used Knowledge distillation to transfer knowledge in Active Learning setting with very few labels.

May 2022 – Jul 2022 · 2 mos · Gurgoan

• I learned and used Graph-based ML and statistical techniques for Card fraud detection.
• We collectively filled a Patent on 'finding unregistered merchants doing crypto transactions through graph-based techniques.'
• I also implemented Label Propagation Algorithm and Breach Radar Algorithm using python and SQL and beat the existing previous results of finding the compromised merchants of the company.

Jul 2021 – Oct 2021 · 3 mos · Ahemdabad

• Worked with Prof Ankit Kariya (IIMA).
• We tried finding bi-grams from transcripts of conference calls of around 2500 public US companies related to a particular field.
• The primary purpose of the research was to find out whether Fed's action impacts companies' upcoming decisions.
• We used https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/FMJDCD, which combines human knowledge and computer power to find out words.
• We did a different search for financial and non-financial companies for better clarity.

Dec 2020 – Jan 2021 · 1 mo · Mumbai, Maharashtra, India

• I learned strategies of market segmentation and their use in customer risk profiling and management which later helped in developing a customer ranking system to improve customer risk profiling and risk management, accommodating the situation caused by the pandemic, leaving company in a position to take better decisions.
• Later designed database schema and developed a master database integrating all the customer attributes for smooth use of companies data of future projects.