Ankit Singh

I am a Senior Research Engineer working on vision-language models, multimodal foundation models, and the Falcon model family. Previously, I was a research student in the Computer Science & Engineering Department at IIT Madras, working on computer vision and deep learning. I completed my undergraduate studies in Computer Science at NIT Silchar.

Email  /  Google Scholar  /  Github  /  Twitter  /  LinkedIn  /  Connect

My research interests lie in computer vision, deep learning, and multimodal AI. My current work focuses on vision-language models (VLMs), early-fusion multimodal architectures, vision foundation models, and label-efficient learning across images and video. I am also interested in video understanding, representation learning, domain adaptation, and efficient model deployment.

• Released Falcon Perception and Falcon-OCR, the world's smallest OCR model (2026)
• One paper at ECCV 2026
• Two papers at CVPR 2026: SigLino and VisRes Bench
• Vision-Language Models Can't See the Obvious at ICCV 2025
• Harnessing Frozen Unimodal Encoders for Flexible Multimodal Alignment at CVPR 2025
• Granted patent on energy-optimized neural networks for in-vehicle road-scene understanding (FR3151115A1, 2025)
• Contributed to the Falcon 3 and Falcon2-11B open model families (2024)
• On permutation symmetries in Bayesian neural network posteriors at NeurIPS 2023
• Institute Research Award 2022, IIT Madras
• Student Volunteer for NeurIPS 2021
• Semi-Supervised Domain Adaptation accepted at NeurIPS 2021
• Semi-Supervised Action Recognition accepted at CVPR 2021
• Mitigating Dataset Imbalance accepted at ECCV-W 2020

Selected publications in reverse chronological order. For the full list, visit my Google Scholar profile.

Reviewer: CVPR, ICCV, ECCV, ICLR, NeurIPS, AAAI, WACV, ACCV, BMVC, TPAMI

Website template from Jon Barron

* denotes equal contribution