Saurabh Sharma

About me

Doctoral candidate in Computer Science at the University of California in beautiful Santa Barbara, specializing in machine learning and network science. I have worked in various areas of machine learning including computer vision (CV), graph ML, natural language processing (NLP) and generative models. My experience spans both academic as well as industrial research in labs at IITB, MPII, UCSB, and Amazon. My publications have been accepted at top ML conferences such as CVPR, ICCV, IJCAI, and the WebConf. I also have experience working on large-scale production systems at Fortune 500 companies such as Goldman Sachs and Microsoft. Outside of work, I can be found playing tabla or MTBing back-country in the Los Padres National Forest.

News

[09/2025] I'm TAing the Introduction to Computational Science course offered in Fall 25 at UCSB.
[04/2025] I attended the Web Conference, 2025 in Sydney, NSW, Australia.
[03/2025] I successfully advanced towards doctoral candidacy for my PhD in Computer Science at UCSB.
[01/2025] Our work on Dynamic Gradient Influencing for Viral Marketing Using Graph Neural Networks is accepted to the Web Conference, 2025.
[12/2024] I attended the Neural Information Processing Systems Conference (NeurIPS), 2024 in Vancouver, BC, Canada.
[08/2024] I TAed the Introduction to Computational Science course offered in Summer 24 at UCSB.
[04/2024] Our work on Dynamic Gradient Influencing for Viral Marketing Using Graph Neural Networks is available as a preprint on ArXiv.
[09/2023] I finished another Applied Scientist internship at Amazon Science, San Diego.
[05/2023] I successfully passed my Major Area Examination towards my PhD in Computer Science at UCSB.
[04/2023] Our work on Learning Prototype Classifiers for Long-Tailed Recognition is accepted to the International Joint Conference on Artificial Intelligence (IJCAI), 2023.
[04/2023] My research is funded by the NSF Agent-Centric Threat Intelligence and Operation (Action) AI Institute.
[02/2023] Our work on Learning Prototype Classifiers for Long-Tailed Recognition is available as a preprint on ArXiv.
[01/2023] I TAed the Introduction to Computer Vision course offered in Winter 23 at UCSB.
[11/2022] I attended the Neural Information Processing Systems Conference (NeurIPS), 2022 in New Orleans, Louisiana.
[09/2022] I finished my Applied Scientist internship at Amazon Science, San Diego.
[04/2022] I TAed the Introduction to Computational Science course offered in Spring 22 at UCSB.
[01/2022] I TAed the Machine Learning course offered in Winter 22 at UCSB.
[10/2021] I TAed the Capstone course for Data Science offered in Fall 21 at UCSB.
[08/2021] I attended the Virtual International Conference on Knowledge Discovery and Data Mining conference (KDD), 2021.
[04/2021] I attended the Virtual Siam International Conference on Data Mining (SDM), 2021.
[10/2020] I started my PhD in Computer Science at the University of California Santa Barbara, Santa Barbara, USA.
[10/2020] I got my MSc in Computer Science from Saarland University, Saarbruecken, Germany.
[09/2020] I presented our work on Long-Tailed Recognition using Class-Balanced Experts at the Virtual German Conference on Pattern Recognition (GCPR) 2020.
[09/2020] Our code for Long-Tailed Recognition using Class-Balanced Experts is released.
[08/2020] Our work on Long-Tailed Recognition using Class-Balanced Experts is accepted to the German Conference on Pattern Recognition (GCPR) 2020.
[10/2019] We presented our work on Monocular 3d human pose estimation by generation and ordinal ranking at the International Conference on Computer Vision (ICCV), 2019 in Seoul, South Korea.
[07/2019] Our work Monocular 3d human pose estimation by generation and ordinal ranking is accepted to the International Conference on Computer Vision (ICCV), 2019.
[06/2019] We presented our work f-vaegan-d2: A feature generating framework for any-shot learning at the Computer Vision and Pattern Recognition (CVPR) conference, 2019 in Long Beach, California, USA.
[03/2019] Our work f-vaegan-d2: A feature generating framework for any-shot learning is accepted to the Computer Vision and Pattern Recognition (CVPR) conference, 2019.

Publications