Ahmad Mahmood

About Me

I am a Machine Learning Engineer at Cast Insights in San Francisco, building production multimodal retrieval and reasoning systems across text, image, and audio, along with the ML infrastructure to serve them at scale.

Previously, I completed my MSc in Computer Science at ETH Zurich (GPA: 5.4/6), where I did my Master’s Thesis at Disney Research Zurich on real-world image deblurring using a novel diffusion-SSM hybrid architecture. Before that, I worked as a research intern at INSAIT (Sofia University), supervised by Dr. Danda Paudel and Dr. Luc Van Gool, where our work on trajectory-controlled video generation was accepted at ICCV 2025. I was also a researcher at the Intelligent Visual Analytics Lab (IVAL) at MBZUAI, supervised by Dr. Salman Khan.

I hold a BSc in Computer Science from the Lahore University of Management Sciences (LUMS) (GPA: 3.8/4), where I graduated in the top 10% with Distinction.

Apart from all that, you will probably find me playing chess during my free time ♕.

Research Interests

• Computer Vision: Image/Video Restoration, Video Generation, Vision-Language Understanding
• Generative Models: Diffusion Models, Score-based Models, State-Space Models
• Machine Learning: LLM Reasoning, Retrieval-Augmented Generation, Neural Network Robustness

News

• [Jul. 2025] Joined Cast Insights as a Machine Learning Engineer in San Francisco.
• [Feb. 2025] Started my Master’s Thesis at Disney Research Zurich on image deblurring using diffusion-SSM hybrid architectures.
• [Sep. 2024] Assisted Dr. Michael Hoffmann in the course Algorithms Lab at ETH Zurich.
• [Mar. 2024] Started my internship as a computer vision researcher at INSAIT under the supervision of Dr. Danda Paudel and Dr. Luc Van Gool.
• [Sep. 2023] Started my MSc in Computer Science at ETH Zurich.
• [June. 2023] Completed my undergraduate degree in Computer Science at LUMS. Graduated in the top 10% with Distinction and Dean’s Honour List.
• [Jan. 2023] Our paper on Boosting Adversarial Transferability Using Dynamic Cues is accepted at the International Conference on Learning Representations (ICLR 2023).
• [Sep. 2022] Assisted Dr. Agha Ali Raza in the graduate level course CS 535 (Machine Learning) at LUMS.
• [June. 2022] Started working as a student researcher under the supervision of Dr. Ihsan Ayyub Qazi and Dr. Agha Ali Raza on robust Federated Learning over non-IID data.
• [Mar. 2022] Started working as a researcher at MBZUAI, supervised by Dr. Salman Khan, on adversarial transferability from image to video models.
• [June. 2021] Became the Vice-President Events of the LUMS Students Mathematics Society (LSMS).
• [Apr. 2021] Joined the CSALT lab at LUMS.
• [July. 2019] Represented Pakistan at the International Mathematical Olympiad (IMO) in Bath, United Kingdom.

Publications

Services

Conference Reviewes

• Asian Conference on Machine Learning (ACML) 2022

Daily Chess Puzzle ♕

Think you can solve it? A new puzzle every day, courtesy of Lichess.

Powered by Jekyll and Minimal Light theme.