Oct 2025 – Jan 2026 · 3 mos · Amsterdam, North Holland, Netherlands

Apr 2021 – Oct 2025 · 4 yrs 6 mos · Amsterdam, North Holland, Netherlands

• Final year PhD student, working on off-policy learning and evaluation from user interactions under the supervision of prof.dr. Maarten de Rijke, and dr. Harrie Oosterhuis.
• PhD thesis is available here: (https://dare.uva.nl/search?identifier=669b7ddf-8c57-44c1-917d-9160ae14c04e).

Aug 2024 – Dec 2024 · 4 mos · Manhattan, New York, United States · On-site

• Worked with Modern Recommendation Systems (Content Understanding team). Proposed LOOP, a novel reinforcement learning-based fine-tuning method for text-to-image diffusion models. Publication from the internship accepted at TMLR 2026.
• Paper: https://openreview.net/forum?id=i8WJhKn455.

Aug 2023 – Dec 2023 · 4 mos · Greater London, England, United Kingdom · On-site

• Research intern with the Instagram Modern Recommender Systems (MRS) team; worked on:
• Large-scale off-policy learning for two-stage recommender system (publication accepted at ICTIR 2025).
• Contributed to a mixture-of-experts model for candidate generator system for Instagram reels recommendation.

Jul 2018 – Jan 2021 · 2 yrs 6 mos · Bengaluru Area, India

• Worked with the search team on various product ranking and query understanding problems:
• Session length prediction for product search (SIGIR'20)
• User clustering using Bayesian models (ECNLP@ACL'20)
• User return time prediction using a temporal point process-based model for early detection of user churn (UMAP 2023)
• Learning to rank to deliver a personalized search experience to the users.
• Short text classification to identify the product category of the query.

Jan 2018 – May 2018 · 4 mos · Bangalore, India

• Worked on the problem of fake news detection, given it's social media interaction.
• Research publication from the project accepted at ASONAM'18, Barcelona.

May 2017 – Oct 2017 · 5 mos · Pune Area, India

• Worked on the problem of adverse drug reaction mention extraction from social media using machine learning and information retrieval techniques.
• Work resulted in publications at NIPS'17 ML4H, CIKM'17 and ECIR'18.

Aug 2016 – Dec 2016 · 4 mos

• Teaching assistant for the machine learning course (CSE577) under Prof. C.V Jawahar.

Dec 2015 – Apr 2016 · 4 mos

• Projects undertaken during internship:
• Implementation of word embedding models (Glove and word2vec) in theano
• Sentiment Analysis system using Recurrent Neural Network for Social Media data using multiple Deep Learning methods including: character based LSTM model, LSTM-CNN model and an ensemble-based model. Demo: http://www.paralleldots.com/sentiment-analysis
• Content-based scientific article recommendation system using deep graph embeddings (work under review at a conference)

Jan 2015 – May 2016 · 1 yr 4 mos · BITS Pilani Pilani Campus

• Worked on the parallelization of the Independent Component Analysis (ICA) algorithm in a hybrid distributed environment using MPI and OpenMP.

Jul 2014 – Dec 2014 · 5 mos · Pilani

• Teaching assistant for Machine learning course for undergrad students.
• Handled lab sessions for Semester 1 2014-15. Lab session was introduced for the first time, designed lab content and syllabus.

Aug 2010 – Apr 2014 · 3 yrs 8 mos

• B.E CSE