SRISHTI YADAV

ELLIS PhD Graduate - University of Copenhagen and University of Amsterdam, member Pioneer Centre of AI and ILLC.

I am a recent graduate and ELLIS PhD Fellow advised by Dr. Serge Belongie (University of Copenhagen) and Dr. Ekaterina Shutova (University of Amsterdam). I am broadly interested in AI and Society. More specifically, I have worked on problems in:

AI Safety and Evaluations
Multimodal Alignment
Risk assessments of LLMs
Tech Governance

Most recently, I was research intern at Google and IBM Research where I worked on safety and governance. Before PhD, I spent couple of years in industry working on ML problems in remote sensing, agri-tech and aerial imagery. In my long past I worked at MILA (bio-diversity monitoring project), Helicopter and VTOL Laboratory - IIT Kanpur and Samsung IoT Innovation Lab.

Collaboration: Like everyone else, always happy to chat about it!

Community Service: I am currently on the board of Women in Computer Vision (WiCV). I was an advisor for WiCV@ICCV2023, WiCV@CVPR 2021, area chair for WiV@CVPR 2020 and on the committee of the WiML@NeuRIPS, 2019.

Note: I am a masters graduate (research-track) from Simon Fraser University, Canada. If you are someone who wants to chat about Masters (research-track) in Canada, DO-NOT hesitate to email me. It’s a competitive but funded program and I’ll be more than happy to tell you about it! I am also happy to answer questions on ELLIS PhD program.

Note: If you are planning to apply for ELLIS PhD program, feel free to send me questions about our lab and/or the program.

Selected Updates:

2026/04/30: 2 papers accepted at ICML. 'When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation' and 'Who Evaluates AI's Social Impacts? Mapping Coverage and Gaps in First and Third Party Evaluations'
2026/01/01: Wrapped up my research internship at Google.
2025/11/06: My internship work with IBM Research on stakeholder-grounded policy distillation pipeline 'Who Sees the Risk? Stakeholder Conflicts and Explanatory Policies in LLM-based Risk Assessment' has been accepted at International AI Governance Workshop (AIGOV), AAAI 2026.
2025/10/12: Our paper 'Multi-Modal Framing Analysis of News' has been accepted to EMNLP 2025 Main.
2025/02/27: Our paper 'Cross-modal Information Flow in Multimodal Large Language Models' has been accepted to CVPR 2025

Rant: I used to have a fancy website but hated dependency on ever changing web frameworks. Back to basics. HTML is cool. Trying to be part of 100kb.club