About Me

Hello! I am a PhD student at the University of Washington advised by Prof. Shwetak Patel as a part of the Ubicomp Lab.

My research interests lie at the intersection of human-computer interaction, systems, and intelligence, with primary application areas being speech/conversational AI and AI for health.

I previously obtained my bachelor’s degree in computer engineering from the University of Michigan ECE, where I was fortunate to work with Prof. Robert Dick on embedded systems and resource-constrained ML.

Contact me at vysri [at] cs.washington.edu!

News

🏆 03/26 Honored to be a recipient of the CIFAR Next Gen Trainee Fellowship!

08/25 Back at Google, this time on the XR & Devices team. Stay tuned for exciting things!

📄 05/25 Our paper "Substance over Style: Evaluating Proactive Conversational Coaching Agents" was accepted to ACL 2025 Main Track! Poster presentation in Vienna, Austria.

🏆 01/25 Received the NCWIT AiC Collegiate Award for my research on human-centered intelligent systems!

🏆 10/24 Won the Madrona Prize Runner Up for my work on Knowledge Boosting during Low-Latency Inference!

07/24 Started a position as a student researcher at Google Health!

📄 06/24 Our paper "Knowledge Boosting: Model Collaboration During Low-Latency Inference" was accepted to Interspeech 2024! Oral presentation in Kos Island, Greece.

Featured Publications

Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice Agents

V. Srinivas^†, Z. Englhardt^†, V. Iyer, S. Patel

^†Co-primary authors

A responsive on-device small model "mouth" speaks while a remote large model "brain" thinks and performs more complicated tasks in the background without real-time constraints.

Substance over Style: Evaluating Proactive Conversational Coaching Agents

V. Srinivas, X. Xu, X. Liu, A. Kumar, I. Galatzer-Levy, S. Patel, D. McDuff

Collaborative conversational paradigms allow better context-seeking and therefore better personalized recommendation in open-ended conversations.

Knowledge Boosting: Model Collaboration During Low-Latency Inference

V. Srinivas, M. Itani, T. Chen, S.E. Eskimez, T. Yoshioka, S. Gollakota

Allowing tiny audio models on-device to receive time-delayed hints from larger off-device models but still output in real time.

Other Publications

The Anatomy of a Personal Health Agent

A.A. Heydari^†, K. Gu^†, V. Srinivas^†, H. Yu, Z. Zhang, Y. Zhang, A. Paruchuri, Q. He, et al.

^†Co-primary authors

An agentic framework for a conversational personal health agent that can understand wearable data, provide coaching, and look through personal health records.

Efficient Video Redaction at the Edge: Human Motion Tracking for Privacy Protection

H. Qiao, V. Srinivas, P. Dinda, R. Dick

An edge-based video redaction system using video & UWB for real-time privacy protection without cloud processing.

LoRa Synchronized Energy-Efficient LPWAN

B. Zhang, V. Srinivas, Y. Zhang, R.P. Dick

A dynamically configuring LoRa network for use in low-power agricultural monitoring.

Vidya Srinivas

News

Featured Publications

Other Publications