Vaibhavi Lokegaonkar
🎧Working on audio generation

Vaibhavi Lokegaonkar

M.S. in Computer Science · University of Maryland, College Park
Multimodal Audio Generation · Inference Optimization

I am Vaibhavi Lokegaonkar, a Master's student in Computer Science at the University of Maryland, College Park. I have been fortunate to work on research under the guidance of Prof. Dinesh Manocha and Prof. Ramani Duraiswami at the GAMMA and PIRL labs.

My research focuses on text- and video-conditioned audio generation. I also have experience with LLM inference optimization through KV-cache management.

I'm open to collaborations, so feel free to drop an email.

News

Publications

WAKMan: ternary plot of KV-cache performance across single-turn, multi-turn and agentic workload mixes
Suhas Kumar Morisetty*, Vaibhavi Lokegaonkar*, Sachin Mathew Varghese, Wesley Harris (*equal contribution)
arXiv · 2026
VIBE: video instruction-aligned background music generation
Aryan Vijay Bhosale*, Vaibhavi Lokegaonkar*, Vishnu Raj, Gouthaman KV, Sreyan Ghosh, Ramani Duraiswami, Lie Lu, Dinesh Manocha (*equal contribution)
EMNLP Findings 2026
Video-Robin overview: text and video conditioned music generation
Vaibhavi Lokegaonkar*, Aryan Vijay Bhosale*, Vishnu Raj, Gouthaman KV, Ramani Duraiswami, Lie Lu, Sreyan Ghosh, Dinesh Manocha (*equal contribution)
NeurIPS 2026 Creative AI Track
SPUR: dataset curation pipeline and model architecture
S Sakshi, Vaibhavi Lokegaonkar, Neil Zhang, Ramani Duraiswami, Sreyan Ghosh, Dinesh Manocha, Lie Lu
arXiv · 2025
MMAU-Pro benchmark example questions
Sonal Kumar, Šimon Sedláček, Vaibhavi Lokegaonkar, Fernando López, Wenyi Yu, Nishit Anand, et al., Sreyan Ghosh, Ramani Duraiswami
AAAI 2026
Chao-Han Huck Yang, Sreyan Ghosh, Qing Wang, Jaeyeon Kim, Hengyi Hong, Sonal Kumar, Guirui Zhong, Zhifeng Kong, S Sakshi, Vaibhavi Lokegaonkar, et al.
DCASE 2025 Challenge
SSBD+ activation heatmaps on video frames
Vaibhavi Lokegaonkar, Vijay Jaisankar, Pon Deepika, Madhav Rao, T K Srikanth, Sarbani Mallick, Manjit Sodhi
IEEE Healthcom 2023

Contact

The best way to reach me is by email at vlokegao@umd.edu. I am happy to discuss research ideas and potential collaborations.