Currently, I am an AI Research Engineer at Anaconda, continuing my AI safety and evaluation research following Anaconda's acquisition of Enkrypt AI in 2026.

I joined Enkrypt AI as its Founding ML Research Engineer, where I designed and evaluated adversarial testing pipelines spanning 250+ foundation models, built production guardrail systems, and studied failure modes of AI models in real-world deployments. This work gave me a unique perspective on applied AI security bridging cutting-edge research with the deployment challenges faced by companies building AI systems today.

Previously, I was a Research Associate at IISc Bangalore, working under Prof. Prathosh A. P.. My work there spanned Neural Machine Translation for Indic languages, machine unlearning in generative models, and privacy-preserving ML. These projects deepened my interest in differential privacy and federated learning as practical approaches for building privacy-aware systems.

I am also collaborating with Fabrizio Frasca (Postdoctoral Researcher, Technion) on a project titled Beyond Text: Exploring Adaptations of LLMs for Graph-Based Tasks. In this work, we aim to investigate how pretrained LLMs can be adapted through parameter-efficient fine-tuning to effectively process relational and graph-structured data, and to study the resulting impacts on invariance, robustness, and generalization to new graph tasks.

News

Selected Publications

Research Projects

Recent Blog Posts

Miscellaneous

I started my research journey in 2020 with text summarization at Helppr AI. Since then, I've worked across language models, unlearning, and generative AI. Today I'm building systems that are robust and scalable in practice. In my free time, I contribute to open source projects and explore the latest developments in AI safety.