Hi, I am Vatsalya

MS, Machine Learning and Data Science, University of California San DiegoB. Tech, Electronics and Communication Engineering,Indian Institute of Technology Bhubaneswar
EmailLinkedInGitHubLinkLink

Research Interests

My major research interest lies in building autonomous systems, especially in planning and control.  The idea that an autonomous system can safely and optimally operate in a noisy ever-changing environment really excites me. The applications of such systems in healthcare, transportation etc. would be immense. As such I want to work with learning-based controls for dynamical systems. I am interested in Reinforcement Learning and Control Theory. I am currently working on Model Predictive Control for continuous-time systems with Continuous Control Barrier Functions to guarantee safety.

Given the recent advancements in LLMs and image generation tools like Stable Diffusion, multi-modal learning has now become the next frontier. GPT-4 has been the first step in this direction. However, they still suffer from hallucinations and scalability issues which become important in production. I am interested in how these tools are developed and can be integrated with other products efficiently to provide better utility to consumers.

Short Bio

I am currently at Google working with the Structured Data Team at Vertex AI, Google Cloud. I am working on integrating text and image embedding models for multi-modal learning in Vertex AI. 

I am also enrolled as an MS student in the Electrical and Computer Engineering Department at University of California San Diego specializing in Machine Learning and Data Science. Before that, I got my bachelor's degree in Electronics and Communication Engineering at the Indian Institute of Technology Bhubaneswar in 2021.

After my undergrad, I was associated with Quriate Education as a founding engineer. There I took up both leadership and development roles establishing the engineering team along with software design and development. The startup currently serves more than 1500 users. Previously, I was in the Global Merhcant Lending Team at PayPal where I developed and mainted large volumne ETL pipelines.

For my bachelor's thesis, I was advised by Dr Siddhartha S Borkotoky . We developed a novel Application Layer Coding schemes for wireless sensor networks/IOT in delay sensitive applications. For brief info about my experiences and projects please look at the work experience and projects page. The details of my work can be found in my CV here.

Publications