Kartik Narayan

I am a 3rd year (final year) Ph.D. student in the Computer Science department at Johns Hopkins University, where I am a member of VIU lab, advised by Dr.Vishal Patel. My current interests lie in multimodal LLMs, agents, long-horizon tool calling, agentic post-training, agentic search, reasoning and video generation. My prior research focused on computer vision and its applications in face analysis, understanding, and recognition, with a particular emphasis on multimodal LLMs.

Currently in my Ph.D., I have been fortunate to intern at:

1. Salesforce AI Research Summer 2026 (Xiangyu Peng, Prafulla Kumar Choubey, Jiaxin Zhang, Silvio Savarese, Chien-Sheng Wu): Developing efficient multimodal agentic search methods using self-distillation to provide intermediate reward signals.
2. Apple Spring 2026 (Tian Cao, Zheng Yang, Peng Zhou, Lunbo Xu, Xiaohui Tao, Zhe Gan): Developed a multi-image, vision-centric benchmark for evaluating multimodal agentic search, and analyzed failure modes of multimodal models in agentic search settings involving multiple visual entities.
3. Apple Summer 2025 (Navid Shiee, Peter Grasch, Chao Jia, Yinfei Yang, Zhe Gan): Equipped multimodal LLMs with web-search tools and enabled reasoning over iterative search calls; developed methods to incentivize self-correction and self-verification following an initial model response.

Prior to my doctoral studies, I worked as an undergraduate researcher under Prof. Richa Singh and Prof. Mayank Vatsa at the Image Analysis and Biometrics (IAB) Lab, IIT Jodhpur, where I worked on deepfake video generation.

Email  /  CV  /  Google Scholar  /  LinkedIn  /  Twitter  /  Github

Open to full-time opportunities starting from December 2026.

My current research interests are multimodal LLMs, agents, long-horizon tool calling, agentic post-training, agentic search, reasoning and video generation. My prior research focused on computer vision and its applications in face analysis, understanding, and recognition, with the goal of developing robust open-world algorithms that can be deployed for real-world impact.

See my Google Scholar profile for the complete and most recent publications.

Template credits. Last updated September 2026.