Saswat Meher

profile_pic.jpg

I’m a Machine Learning Engineer at Samsung Research Bangalore, building multimodal (text + image) retrieval systems. I earned my M.Tech in CSE from IIT Bombay, where I worked with Prof. Abir De and Prof. Soumen Chakrabarti, and a B.Tech in CSE from NIT Rourkela.

My master’s work centered on neural architectures for (sub)graph retrieval—basically teaching models to understand and fetch the right pieces of graph-structured data.

More broadly, I’m excited about representation learning for retrieval across modalities: text, images, and graphs.

selected publications

  1. position.png
    Position: Graph Matching Systems Deserve Better Benchmarks
    Indradyumna Roy, Saswat Meher, Eeshaan Jain, and 2 more authors
    In Forty-second International Conference on Machine Learning Position Paper Track, 2025
    HTML
  2. graphedx.png
    Graph Edit Distance with General Costs Using Neural Set Divergence
    Eeshaan Jain, Indradyumna Roy, Saswat Meher, and 2 more authors
    In Advances in Neural Information Processing Systems, 2024
    HTML Code
  3. algal.jpg
    Predicting algal biochar yield using eXtreme Gradient Boosting (XGB) algorithm of machine learning methods
    Abhijeet Pathy, Saswat Meher, and Balasubramanian P
    Algal Research, 2020
    DOI