Hi! I am a final year PhD student at UC Irvine, where I’m affiliated with Learning, Inference & Vision Group and advised by Prof. Erik Sudderth. My PhD is generously supported by an HPI fellowship. I interned at Amazon.com Services LLC in 2020.
My research interests encompass machine learning, computer vision, and probabilistic reasoning. I am dedicated to enhancing the reliability and robustness of generative models for specific user applications, building on foundational concepts such as probabilistic modeling and optimization. My PhD thesis is centered on advancing the consistency and performance of image generative models, including Diffusion Models and Hierarchical VAEs for image inverse problems like image inpainting, deblurring etc. Specifically, I focus on adapting and aligning latent representations to ensure that the generated images accurately adhere to user observations.
Before graduate school, I was a Budding Scientist at Conduent Labs (formerly known as Xerox Research Centre India). I received a Bachelors in Electrical Engineering (Instrumentation) with a minor in Computer Science Engineering from Indian Institute of Technology, Kharagpur in 2017. During that period, I was also a visiting researcher at Duke University and an intern at TCS Innovation Labs.
When I’m not working, I like to spend my time cooking, dancing, playing tennis, learning about cultures and taking short trips outside the city!
You can reach me at sakshia1@uci.edu or visit my LinkedIn profile. A short description of my works is present in my CV.