Published numerous papers in top AI conferences and journals. Recent areas include robust RLHF, evaluation of LLM alignment with theoretical guarantees, uncertainty quantification for LLMs, differential privacy for synthetic data generation, and agentic workflows for democratizing causal inference tools.
Research Experience
Currently a Senior Research Scientist and Manager at the MIT-IBM Watson AI Lab, leading a team focused on fundamental AI/LLM research collaborating with MIT. He has served as principal investigator for many multi-year research grants with MIT.
Education
PhD in Signal Processing from the University of Michigan in 2017; Postdoctoral research fellow at the Harvard University Statistics department before joining IBM Research.
Background
Research interests include generative AI (primarily LLMs), and making AI more reliable and robust, enabling an increasing level of real-world autonomy for machines. Keywords: Large language models, uncertainty quantification, language model alignment/RL, optimal transport, causal inference, signal processing.
Miscellany
Contact: Email: kristjan.h.greenewald@ibm.com; Address: 314 Main St, Cambridge, MA, IBM Research, MIT-IBM Watson AI Lab