Ashutosh Trivedi

Ashutosh Trivedi is an Associate Professor of Computer Science at the University of Colorado Boulder, where he leads research in the Programming Languages and Verification (CUPLV) group. His work focuses on making AI systems more trustworthy through the application of formal methods, particularly in safety-critical applications like medical devices and legal software.

Trivedi's research interests span Safety in AI, Reinforcement Learning, Formal Methods, Software Fairness, and Software Accountability. He uses mathematical precision from formal methods to address the challenges of AI systems that learn from data rather than following fixed rules. His approach involves using formal languages, automata, and logic to turn vague natural-language instructions into clear specifications, enabling collaborative programming with AI that is explainable and reliable.

His recent publications reveal a strong trend toward bridging formal verification techniques with reinforcement learning, particularly in safety-critical domains. His work frequently appears in top venues including AAAI, CAV, ICSE, and NeurIPS, with a focus on making AI systems behave safely, fairly, and responsibly even as they learn and adapt.

• Distinguished Paper Award at CAV for Regular Reinforcement Learning (2024)
• Senior Member of the ACM (2024)
• Royal Society Wolfson Visiting Fellowship (2024)
• NeuS 2025 Disruptive Idea Award
• NeuS'25 paper accepted for oral presentation

Trivedi actively mentors PhD students, with recent graduates including Shadi Tasdighi Kalat (2025), Mateo Perez (2025), John Komp (2024), Vishnu Murali (2024), and Taylor Dohmen (2024). His students' theses focus on formal languages for reinforcement learning, safety verification, and principled AI system design.

He leads the Programming Languages and Verification (CUPLV) research group at the University of Colorado Boulder, which develops formal techniques to verify and improve the reliability of programming languages and software systems, with particular emphasis on AI safety and accountability.