Scholar
Tejaswi Kasarla
Google Scholar ID: e_mkTxMAAAAJ
PhD Candidate, University of Amsterdam
Citations & Impact
Citations
110
H-index
3
i10-index
2
Publications
10
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Publications: 'Balanced Hyperbolic Embeddings are Natural Out-of-Distribution Detectors', Preprint 2025; 'Hyperbolic Safety-Aware Vision-Language Models' accepted to CVPR 2025 as a Highlight; 'Maximum Separation' accepted to NeurIPS 2022; 'Region Based Active Learning for Efficient Labelling in Semantic Segmentation' accepted to WACV 2019. Organized Events: Women in Computer Vision (WiCV) Workshop, CVPR 2021 and 2022; WiCV Social, ICCV 2021. Reviewer: ICCV 2023, WiCV Workshop @ CVPR 2023, NeurIPS 2023, ICLR 2024.
Research Experience
- Research Scientist Intern at Meta FAIR, Paris, June 2025 - December 2025; Teaching Assistant for Foundation Models (FoMo) course at UvA, April 2025; Poster presentation at ELLIS Doctoral Symposium 2022, September 2022; Poster presentation at Summer School of Machine Learning at Skoltech (SMILES), August 2020; Computer Vision Researcher at Research and Technology Center, Bosch, May 2019 - October 2021; Intern at Research and Technology Center, Bosch, June 2018 - October 2018.
Education
- PhD: VIS Lab, University of Amsterdam, Supervisors: Pascal Mettes and Rita Cucchiara, October 2021 - Present; Master's Thesis Defense: June 2019.
Background
- Research Interests: Non-Euclidean representation learning (hyperspherical and hyperbolic deep learning) for open-world understanding and multimodal foundation models. Professional Area: Computer Vision, Machine Learning.
Miscellany
- Hobbies: Photography, Home Barista, Exploring the specialty coffee scene wherever she goes.
Co-authors