Work

Research, publications, and patents

Experience

Research Scientist · Dolby Laboratories
October 2021 – Present · Sunnyvale, CA

Focusing on generative video models and video restoration with multimodal Diffusion models.

Software Engineer Lead · Citigroup
June 2016 – December 2019 · Chennai, India

Led automation infrastructure engineering for enterprise-scale digital payment systems across APAC and EMEA. Architected and deployed CI/CD pipelines, built production Spring Boot applications serving 200+ developers, and engineered scalable test automation frameworks with Java, Selenium, and Maven. Managed a 7-person cross-functional team certifying 10+ concurrent projects for Citibank's Digital Payments & Settlements division. Designed and deployed automation infrastructure for Apple Pay integration, reducing test execution time by 75% and data preparation time by 90%. Technologies: Spring Boot, Java, Jenkins, Maven, Selenium, REST APIs, MySQL, Docker, Git.

Publications

Spatio-Temporal Film Grain Removal in Video Using Attention-Guided Neural Networks
S. H. Musunuri, A. Choudhury, G. M. Su
IEEE 8th International Conference on Multimedia Information Processing and Retrieval (MIPR), 2025
IEEE Xplore
Deep Restoration of Archival Videos: Developments, Challenges, and Opportunities
S. Gehlot, S. H. Musunuri, S. Bera, A. Kar, G. M. Su
International Conference on Computing, Networking and Communications (ICNC), 2025
conf-icnc.org
Event-Based Video Frame Extrapolation with Recurrent Neural Networks
S. H. Musunuri, A. Choudhury, G. M. Su, S. Park
58th Asilomar Conference on Signals, Systems, and Computers, 2024
IEEE Xplore
AE-StyleGAN: Improved Training of Style-Based Auto-Encoders
S. H. Musunuri, L. Han, M. R. Min, R. Gao, Y. Tian, D. Metaxas
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022
DOI: 10.1109/WACV51458.2022.00103

Patents

Trim Pass Metadata Prediction in Video Sequences Using Neural Networks
S. H. Musunuri, S. S. Rotti, A. K. A. Choudhury
World Intellectual Property Organization, Patent No. WO2023224917A1, 2023
View on Google Patents