From a village in Karnataka to a research career in Munich

I was born in Neeralagi, a small village in Karnataka, India, where access to computers was rare and the nearest path into technology ran through years of disciplined study. That path took me to the Siddaganga Institute of Technology in Tumkur for my M.Tech in Computer Science, and from there to the Research and Technology Centre at Robert Bosch in Bangalore, where I first worked on extracting structured insight from unstructured text — concept detection and clustering from news feeds. That project planted the question that still drives my research today: how can a machine understand what a document is actually saying?

In 2016, I moved to Germany to pursue an M.Sc. in Computer Science at the University of Passau. It was a leap across more than geography — a new language, a new academic culture, and a much harder set of problems. My master's thesis on open information extraction sharpened a focus on natural language processing that has shaped everything since. A short stint as an AI Engineer at the Digital Product School, working alongside IBM Watson in Munich, and later as a Machine Learning Engineer at Rfrnz GmbH in Berlin, gave me an industry perspective I still draw on: research only matters if someone can use it.

In 2019, I returned to Passau as a doctoral researcher, drawn to one of the hardest document-understanding problems there is: patents. Patent text is dense, adversarial, and deliberately ambiguous — written to be defensible, not readable. Teaching Deep Learning, Information Retrieval, and Web Science to Passau students while supervising their thesis projects taught me as much as the research itself. Since 2023, I've continued this work as a Ph.D. candidate and research assistant at Hochschule München, building AI algorithms that automate prior-art search and surface the novelty hidden inside millions of patent documents.

That work — recognized with a Top 6 finish at the EPO Codefest — sits at the intersection of information retrieval, large language models, and explainable AI. As I head toward completing my Ph.D. by the end of 2026, my research is widening beyond patents toward a broader goal: building domain-specific AI systems that let any organization, from patent offices to hospitals to manufacturers, retrieve and trust the knowledge locked inside their own documents.