CV

Warner Bros. Discovery (HBO) · San Francisco, CA

Leading large-scale video understanding and content moderation systems, including automated harmful content detection and multi-modal safety classifiers for HBO's streaming platform. Developing temporal segmentation, scene/shot boundary detection, VLM-powered video indexing, and LLM-enhanced metadata generation for fine-grained video search and discovery across Warner Bros.' catalog.

Tubi · San Francisco, CA

Developed AI-driven content analysis and personalization solutions using Vision-Language Models, including brand identification and logo detection, scene segmentation, personalized trailers and content artwork generation through user preference modeling, and LLM-powered recommendation enhancements.

Aibee U.S. Corporation · Palo Alto, CA

Designed and implemented CV models for human tracking in crowds, person re-identification, and AI-based benchmarking tools for tracking.

Netflix · Los Gatos, CA

Developed a semi-supervised activity detection system for identifying events in movies without human annotations and trained an attentive deep neural network to optimize movie billboard design based on visual patterns.

Nielsen · Orlando, FL

Developed automated methods for TV advertisement analysis focusing on classification, tagging, and descriptive analysis of broadcast ads.

Center for Research in Computer Vision (CRCV), UCF · Orlando, FL

PhD research with Prof. Mubarak Shah. Topics include Visual Question Answering, Visual Text Correction, video generation, video retrieval, and multimodal AI.

University of Central Florida · FL, USA

Advisor: Prof. Mubarak Shah, CRCV
Dissertation: Video Content Understanding Using Text

University of Central Florida · FL, USA

Sharif University of Technology · Tehran, Iran

Languages: Python, C/C++, SQL, MATLAB; familiar with Scala, Java

ML Frameworks: PyTorch, TensorFlow, Keras

VLMs & LLMs: Expert in state-of-the-art VLMs and LLMs; instruction tuning, multimodal fine-tuning for video/content understanding

Cloud & Big Data: AWS, Apache Spark, PySpark

DevOps: Docker, CI/CD pipelines

Machine learning techniques for advanced frequency management

U.S. Patent · US20240314371A1

Multimedia scene break detection

U.S. Patent · US20240357217A1