I occasionally write about things I'm thinking through. You can find my posts here.

I design AI systems that understand abstract, subjective concepts — like atmosphere, brand identity, or sustainability messaging — in large-scale multimodal data. I'm currently a Postdoctoral Researcher at the University of Amsterdam's Research Center for Sustainable Investments and Insurance (RCSII), where I apply these methods to real business problems in sustainable investments and insurance. I received my PhD on analyzing abstract dimensions in online social multimedia from the University of Amsterdam.

I have applied this work to challenges such as capturing abstract signals like "atmospherics" from user-contributed restaurant images and analyzing how companies communicate sustainability-related messages through high-level visual content.

I specialize in the full ML pipeline: from collecting and cleaning messy, real-world multimodal data to training and deploying models in production. My toolbox includes Python, PyTorch, large-scale vision-language models, and cloud-based infrastructure for scalable deployment. I'm also the co-creator of Exquisitor, an interactive visual search and exploration system that combines conversational queries with relevance feedback to explore millions of images and videos at scale.

I'm currently exploring opportunities where I can apply my expertise in multimodal AI to high-impact problems, whether in industry, research, or at the intersection of both.

Get in touch: contact@ujjwal.nl · LinkedIn · Google Scholar