Angira is a Dphil (PhD) student in Computer Science at the University of Oxford, supervised by Prof. Anisoara Calinescu and Prof. Philip Torr. Her research focuses on bridging the gap between the intended design of AI systems and their actual behavior, with the aim of promoting the reliable application of AI for societal benefit. She works on developing robust and interpretable AI models to mitigate short-term risks and long-term existential threats. A central theme of her research is aligning AI models with evolving human goals, preferences, and values while ensuring their robustness to distributional shifts, particularly in human-in-the-loop scenarios.

Formerly, Angira held leadership roles in projects spanning diverse areas, including natural language processing for legal applications, federated learning in digital twins for supply chains, and robust image segmentation techniques. She has collaborated with the University of Cambridge, Fortune 500 companies, startups, and had brief stints at the Alan Turing Institute and OpenAI. She holds an MSc in Computer Science from Oxford.