Graham Neubig is an Associate Professor at the Language Technologies Institute (LTI) within Carnegie Mellon University (CMU). His research focuses on advancing artificial intelligence, particularly in natural language processing (NLP), multimodal reasoning, and large language models (LLMs). He explores topics such as AI safety, generative AI, and human-AI interaction, with an emphasis on practical applications like machine translation and web-agent systems. His work often involves developing frameworks for evaluating AI systems, such as OpenAgentSafety and BehaviorBox, which assess real-world agent performance and model behavior. Neubig's research also delves into improving LLM capabilities through reasoning analysis, hallucination detection (e.g., ZINA), and culturally aware systems (e.g., CAIRe). He has contributed to open-source projects like Pangea (a multilingual LLM) and frameworks such as Cmulab for model deployment. His recent work addresses challenges in agentic tasks, self-improving agents (Skillweaver), and benchmarking across domains like visual reasoning (VisualPuzzles) and software engineering. Notable achievements include advancing evaluation methodologies for LLMs, developing tools for ethical AI, and creating benchmark suites that test systems under realistic conditions. His lab collaborates on projects like the BrowserGym ecosystem and OpenHands platform, which aim to standardize web-agent research and AI-driven software development. Neubig's contributions span theoretical advancements and practical implementations, bridging the gap between cutting-edge research and real-world applications. He advises students such as Apurva Gandhi and actively publishes in top venues, addressing topics from instruction-following improvements to the societal impacts of AI. His work frequently emphasizes the importance of transparency, controllability, and cultural awareness in AI systems.
Dr. Min Xu is a Courtesy Professor in the Computational Biology Department within the School of Computer Science at Carnegie Mellon University. His research focuses on advancing computer vision and machine learning for biomedical image analysis, particularly cellular cryo-electron tomography (Cryo-ET) and automated science video analysis. He leads a lab developing cutting-edge computational tools for structural biology and medical imaging. Key research directions include: High-resolution 3D Cryo-ET image analysis AI-driven medical image segmentation Few-shot learning for cryo-EM analysis Video analysis frameworks for laboratory automation Notable contributions include the AITom toolkit for Cryo-ET analysis and pioneering work in adapting foundation models for medical imaging tasks. His work has been published in top venues like CVPR, MICCAI, and Nature-associated journals. No academic awards or grants are explicitly listed in the provided text. He maintains an active lab focused on translating computational methods into impactful biomedical research tools.
Chinmay Kulkarni is an Assistant Professor at Carnegie Mellon University's Human-Computer Interaction Institute, where he leads the Expertise@Scale lab. His research integrates large-scale data and automation to transform learning, work, and mentoring systems. Education : Ph.D. in Computer Science from Stanford University (recipient of the Arthur P Samuel Award) Previous Affiliations : Microsoft Research, Barcelona Supercomputing Center His research spans: Human-Computer Interaction design for massive collaboration Voice-controlled interfaces and AI tools Future of work in remote/hybrid environments Behavioral economics through tech interventions Creative entrepreneurship support systems Algorithmic feedback in education Recent publications with AI and education focus show strong trends in voice technology, peer feedback mechanisms, and scalable learning platforms. His lab's systems have been used by >100,000 users across 150 countries. Scientific Awards : Arthur P Samuel Award (Stanford thesis award) Advising & Grants : NSF grant recipient US Department of Education funding Office of Naval Research support Departmental fellowship Labs : Directs Expertise@Scale lab developing systems adopted by Coursera and edX. Current research group includes PhD students Yasmine Kotturi, Julia Cambre, Pranav Khadpe and Masters student Sayan Chaudhry.
Joel Greenhouse is a Professor of Statistics at Carnegie Mellon University (CMU), affiliated with the Department of Statistics & Data Science. He has been on the faculty since 1983 and held leadership roles, including serving as Associate Dean of the College of Humanities and Social Sciences from 1997 to 2002. He also holds an adjunct appointment as Professor of Epidemiology and Psychiatry at the University of Pittsburgh. His expertise spans statistical methodology, clinical trial design, and meta-analysis, with a focus on integrating data from multiple sources to address complex healthcare and public health challenges. Greenhouse earned his Ph.D. in Biostatistics from the University of Michigan and completed a postdoctoral fellowship at CMU. His research emphasizes developing statistical tools for observational studies, clinical trials, and meta-analytic frameworks, particularly in neurology, mental health, and public policy contexts. Notable contributions include analyzing the impact of media on youth suicide rates, improving aphasia classification through automated speech analysis, and evaluating highway safety through driver health data. Education: Ph.D. in Biostatistics, University of Michigan Affiliations: Adjunct Professor at University of Pittsburgh, Member of National Academy of Sciences’ committees Professional Service: Data and safety monitoring boards for NIH/VA studies, co-chair of Federal Motor Carrier Safety Administration review panels His awards include CMU’s Doherty Award for Education, Ryan Teaching Award, and E. Dunlop Smith Award for teaching excellence. His work bridges theoretical statistics with real-world applications, particularly in interdisciplinary collaborations across medicine, psychology, and public policy. Greenhouse’s recent articles highlight trends in leveraging large datasets for clinical insights (e.g., aphasiaBank), re-evaluating environmental and behavioral health associations, and advancing causal inference methods. His interdisciplinary approach ensures statistical rigor addresses societal challenges, from suicide prevention to highway safety.
Tom Mitchell is the Fredkin Professor of AI and Learning and Director of the Center for Automated Learning and Discovery (CALD) at Carnegie Mellon University's School of Computer Science. His research focuses on machine learning, computational neuroscience, and their applications in neuroimaging and natural language processing. He is renowned for pioneering work in developing algorithms to decode brain activity and for contributions to foundational machine learning theory, including co-training and explanation-based learning. Mitchell authored the seminal textbook *Machine Learning* (McGraw Hill, 1997) and has led projects like Never-Ending Learning (NELL), an AI system that autonomously learns from web content. His work bridges computer science and cognitive science, exploring how machines can learn from data and human interaction. Notable research interests include brain-computer interfaces, automated knowledge extraction, and ethical AI. Mitchell's publications span influential journals like *Science* and *Nature*, and he has been recognized for advancing interdisciplinary research in AI and neuroscience. He has advised numerous students and contributed to initiatives like the AAAI Presidential Address on AI and brain sciences. Mitchell's current projects include studying the neural basis of language and developing AI tools for education and healthcare.
Joseph E. Gonzalez is an Associate Professor in the Department of Electrical Engineering and Computer Science at the University of California, Berkeley, co-director of the Sky Computing Lab and RISE Lab, and member of the Berkeley AI Research (BAIR) group. His work bridges artificial intelligence and data systems with significant impact on large language model infrastructure and deployment. His research focuses on large language models (LLMs) including tool use, RAG, and agent systems; LLM deployment infrastructure; edge-based machine learning; cloud computing innovations; and computer vision applications. He addresses the full machine learning lifecycle from training to serving, emphasizing real-time decision systems and secure execution environments. Scientific Awards: Okawa Research Grant NSF Expedition Award NSF CAREER Award Professor Gonzalez mentors 17 current graduate students and numerous former students/post-docs across AI and systems research. His work is funded by the NSF Expedition grant for the RISE Lab, an NSF CAREER Award, and industrial sponsors including major technology companies. He co-directs the RISE Lab advancing real-time intelligent secure execution systems, and the Sky Computing Lab pioneering cloud computing abstractions. These initiatives tackle low-latency systems, online learning algorithms, and security frameworks for intelligent decision-making in physical environments.
Maarten Sap is an Assistant Professor at Carnegie Mellon University's Language Technologies Institute with a courtesy appointment in the Human-Computer Interaction Institute. He also holds a part-time research scientist position at the Allen Institute for AI (AI2) as an AI safety lead. Current affiliations: CMU (2022–present), AI2 (2022–present) Prior: Postdoctoral Researcher at AI2 (2021–2022), Research Intern at AI2 (2018–2019) and Microsoft (2019) His research focuses on enhancing AI systems with social intelligence and addressing social biases in language technology. Key themes include: Ethical AI and Human-Centric Design Narrative Dynamics and Social Context Analysis AI Agents and Social Intelligence Toxic Language Detection and Cultural Bias Mitigation Recent publications examine: AI safety frameworks like HAICOSYSTEM Clinical reasoning alignment (ALFA) Multilingual moderation (PolyGuard) Cultural sensitivity in non-verbal AI (Mind the Gesture) Personality shaping in LLMs (BIG5-CHAT) Scientific Recognition: 2025 Okawa Research Grant Best Paper Runner Up - NAACL 2025 Outstanding Paper - EMNLP 2023 Best Paper - FAccT 2023 Best Paper - WeCNLP 2020 He advises a diverse group of PhD students across CMU and MIT, and has served on multiple program committees including ACL, EMNLP, and FAccT. His work appears in top venues like Nature Machine Intelligence, PNAS, and ACL.
Sean Welleck is an Assistant Professor at Carnegie Mellon University's School of Computer Science, specifically within the Language Technologies Institute (LTI). He leads the L3 Lab and serves as an advisor for the AI for Math Fund. His academic journey includes a PhD from New York University under Kyunghyun Cho and postdoctoral positions at the Allen Institute for Artificial Intelligence and the University of Washington with Yejin Choi. Dr. Welleck's educational background shows a strong foundation in computer science. He earned his PhD in Computer Science from New York University, where he worked under the mentorship of Kyunghyun Cho and Zheng Zhang. Prior to this, he completed his MSE and BSE in Computer Science from the University of Pennsylvania, demonstrating a long-standing commitment to the field. Dr. Welleck's research focuses on bridging informal and formal reasoning with AI, with particular emphasis on developing learning, inference, and evaluation algorithms for large language models. His work spans multiple cutting-edge areas including mathematical reasoning , code generation , inference algorithms , and AI reasoning agents . A significant portion of his recent work involves combining AI with formal methods for mathematics, where he has developed frameworks like Llemma (an open-source language model for mathematical reasoning) and meta-generation (for inference-time algorithms). His research is characterized by a strong theoretical foundation coupled with practical applications that push the boundaries of what AI systems can achieve in formal reasoning domains. Analysis of Dr. Welleck's recent publications reveals a clear research trajectory focused on enhancing language models' capabilities in formal reasoning and mathematical problem-solving. His work demonstrates an evolution from foundational research in neural text generation to increasingly sophisticated approaches that integrate formal methods with deep learning. Key trends include the development of inference-time algorithms that improve model performance without additional training, frameworks for mathematical reasoning that connect informal and formal proofs, and novel evaluation methodologies for language models. His publications consistently appear in top-tier conferences including NeurIPS, ICLR, ICML, and ACL, reflecting the high impact of his contributions to the field. Dr. Welleck's scientific achievements have been recognized with several prestigious awards: NAACL 2025 Best Paper Award ICLR 2025 Oral Presentation (Top 2%) ICLR 2025 Spotlight Presentation (Top 5%) NeurIPS 2021 Outstanding Paper Award (Top 0.1%) for MAUVE NVIDIA AI Labs Pioneering Research Award (2017 and 2018) As an educator and mentor, Dr. Welleck actively guides the next generation of AI researchers. He currently advises multiple PhD students including Pranjal Aggarwal, Weihua Du, Andre He, and Seungone Kim (some co-advised with other faculty), along with MS students Riyaz Ahuja, Jiewen Hu, Qinyue Tan, and Thomas Zhu, and undergraduate Tate Rowney. At CMU, he teaches advanced courses such as Neural Code Generation and Advanced NLP, and has previously taught at New York University and the University of Washington. His commitment to education extends to creating resources like the Thesis Review Podcast and developing tutorials on neural theorem proving that have been presented at major conferences. Dr. Welleck leads the L3 Lab at CMU, which focuses on the intersection of language, learning, and logic. The lab brings together students and researchers to tackle challenging problems in AI reasoning, with particular emphasis on mathematical reasoning and code generation. Recent initiatives include the development of Llemma, an open-source language model specialized for mathematical reasoning, and work on inference-time algorithms that enable language models to improve their performance through additional computation during inference rather than through additional training.
S. Mohadeseh Taheri-Mousavi is an Assistant Professor in the Department of Materials Science and Engineering at Carnegie Mellon University (CMU), part of the College of Engineering. She joined CMU in September 2022 after postdoctoral appointments at MIT and Brown University. Her research is supported by major grants from NASA STRI, DARPA, the Army Research Laboratory, and the Naval Nuclear Laboratory, and she is affiliated with the NextManufacturing Center and the Wilton E. Scott Institute for Energy Innovation. Her educational background includes a Ph.D. from EPFL, Switzerland, and M.Sc. and B.Sc. degrees from Sharif University of Technology, Iran. She was awarded both early and advanced Swiss National Science Foundation fellowships during her postdoctoral studies. Taheri-Mousavi’s research focuses on the intersection of materials science, mechanical engineering, and computer science. She develops multi-scale computational models and AI-driven frameworks—such as AlloyGPT and generative AI agents—to design next-generation structural alloys, particularly for additive manufacturing and extreme environments. Her work emphasizes materials sustainability, industrial decarbonization, and uncertainty quantification in alloy design. The integration of machine learning with Integrated Computational Materials Engineering (ICME) and CALPHAD methods enables rapid exploration of high-dimensional composition and processing spaces. Her recent publications (2023–2025) show a strong trend toward AI/ML applications in alloy discovery, hydrogen embrittlement modeling, and high-temperature aluminum and tungsten alloys. These works reflect a deep commitment to accelerating materials innovation through human-AI collaboration and smart experimental validation. Her scientific honors include prestigious Swiss National Science Foundation fellowships. She has also received seed funding from the Scott Institute for Energy Innovation to study hydrogen embrittlement. She advises a dynamic team of doctoral students and a postdoctoral researcher, working on topics including hydrogen embrittlement, generative AI for welding, and gradient alloys. Her research is funded by high-impact grants from NASA, DARPA, the Army, and the Naval Nuclear Laboratory, supporting transformative projects in structural alloy design. She leads the Taheri-Mousavi Group, which operates within CMU’s Materials Characterization Facility and the NextManufacturing Center. The group focuses on developing novel AI-integrated computational frameworks to guide efficient and intelligent experimentation in alloy development.
David P. Woodruff is a Professor in the Department of Computer Science at Carnegie Mellon University, part of the Theory Group within the School of Computer Science. He is actively involved in academic leadership roles, including chairing the CATCS (Conference on Theoretical Computer Science) and serving as PC chair for SODA 2024 and ICALP 2022. His research focuses on algorithms, data streams, machine learning, numerical linear algebra, sketching, and sparse recovery. He has been recognized with awards such as the Herbert Simon Award for teaching and the PODS Best Paper Award. Woodruff has advised numerous students and postdocs, including notable scholars like Ainesh Bakshi, Rajesh Jayaram, and Hongyang Zhang. His work often addresses foundational challenges in theoretical computer science, with contributions to distributed computing, streaming algorithms, and privacy-preserving techniques. He has published extensively in top conferences like NeurIPS, ICML, FOCS, and STOC, covering topics ranging from low-rank approximation to adversarial robustness in data streams. His teaching includes courses like Algorithms for Big Data and core algorithms courses, reflecting his commitment to both research and education. Collaborations span academia and industry, with applications in genomics and secure computation. Woodruff is a key contributor to the Foundations of Data Science program at the Simons Institute.
Andrew O. Arnold is a Principal Applied Machine Learning Engineer at Shopify and an Adjunct Professor at New York University's Tandon School of Engineering, Department of Finance and Risk Engineering. He earned his Ph.D. in Machine Learning from Carnegie Mellon University and a BA in Computer Science and Artificial Intelligence from Columbia University. Education Ph.D., Machine Learning, Carnegie Mellon University BA, Computer Science and Artificial Intelligence, Columbia University His research focuses on robust machine learning , developing models that perform well in low signal-to-noise regimes, handle distributional shifts (transfer learning), and extract features from unstructured data. Key applications include time series analysis and natural language processing in financial and other domains. Recent publications highlight work on large language models (LLMs) for code generation, including multitask pretraining, contrastive learning, and quantization techniques for efficiency. He has contributed to understanding model robustness and adapting NLP methods to dynamic market conditions. Arnold teaches NYU FRE GY 7871: News Analytics and Machine Learning , covering NLP and ML techniques for quantitative trading strategies. The course emphasizes practical applications of sentiment analysis, text relevance, and novelty detection in financial contexts. He has led teams at Amazon Web Services (AI Labs), served as Chief Scientist at Oracle Alpha, and worked at Microsoft Research, IBM Research, and other institutions. His technical expertise spans code generation , anomaly detection , and NLP for commerce , with patents in these areas.
Jamie Callan is a Professor at Carnegie Mellon University's Language Technologies Institute (School of Computer Science), where he leads research in Information Retrieval and Neural Search Architectures . He teaches advanced courses on search engine design and mentors students in multiple programs. Research Focus: Federated retrieval, knowledge graph integration in search, ClueWeb dataset development, and neural approaches to document ranking Leadership: Past SIGIR Treasurer/Chair, Co-founding Editor of Foundations and Trends in IR, former TOIS Editor-in-Chief His recent work explores: Neural Retrieval: Latent vocabulary for sparse systems, hypothetical documents for dense vector retrieval Dataset Innovation: Maintenance and distribution of ClueWeb09, ClueWeb12, and ClueWeb22 datasets Search Efficiency: Selective search architectures with 90% reduced computational costs Scientific Recognition: International ACM SIGIR Conference Leadership Co-founding Editor-in-Chief, Foundations and Trends in IR Former Editor-in-Chief of ACM TOIS Dr. Callan's Lemur Project has produced Indri/Galago search engines and supported TREC evaluations through dataset contributions.
Han Zhao is an Assistant Professor in the Department of Computer Science at the University of Illinois Urbana-Champaign (UIUC), affiliated with the Department of Electrical and Computer Engineering. He is also an Amazon Scholar at Amazon AI and Search Science. Prior to UIUC, he was a machine learning researcher at D.E. Shaw & Co. Zhao holds a Ph.D. from Carnegie Mellon University's Machine Learning Department, an MMath from the University of Waterloo, and a BEng from Tsinghua University's Computer Science Department. His research focuses on trustworthy machine learning, emphasizing transfer learning (domain adaptation, generalization, multitask/meta-learning), algorithmic fairness, and probabilistic circuits. Applications span natural language processing, signal processing, and quantitative finance. He aims to develop robust, fair, and interpretable ML systems. Recent work includes advancements in domain adaptation theory, multi-task learning optimization, and fair classification post-processing. He advises numerous PhD and master’s students across CS and ECE, co-advising some with colleagues like Hari Sundaram and Ilan Shomorony. Courses taught include CS 442 (Trustworthy ML) and CS 446 (Machine Learning). Key contributions include the MDAN framework for multi-source domain adaptation and theoretical analyses of invariant representation learning. His work balances foundational theory with practical applications, addressing challenges like hyperparameter sensitivity and scalable influence functions.
Jack Mostow is a Research Professor affiliated with the Human-Computer Interaction Institute at Carnegie Mellon University. His work focuses on educational technology, artificial intelligence, and natural language processing. He is renowned for developing Project LISTEN's automated reading tutor that listens to children's oral reading, integrating speech recognition and machine learning to improve literacy. His research addresses challenges in child literacy, automated assessment, and the application of AI in education. Key contributions include advancing speech recognition accuracy for children’s voices, designing adaptive tutoring systems, and exploring EEG and visual features to gauge learning states. He has contributed to global education initiatives, such as the XPRIZE field study, and developed tools for iterative improvement of intelligent tutoring systems. His work emphasizes scalable educational technologies for under-resourced regions and data-driven approaches to enhance learning outcomes. Mostow’s publications span over two decades, addressing topics like reinforcement learning for instructional policies, semi-supervised affect perception, and difficulty-controllable question generation. His research bridges AI, education, and human-computer interaction, aiming to create systems that adapt to learners’ needs while advancing computational methods for educational analytics.
Shixiang (Woody) Zhu is an Assistant Professor in Data Analytics at the Heinz College of Information Systems and Public Policy, Carnegie Mellon University. He holds a PhD in Machine Learning from Georgia Institute of Technology (2022) and B.S./M.S. in Computer Science from Beijing University of Posts and Telecommunications (2017). His research bridges machine learning, operations research, and statistics, focusing on sequential modeling, human-AI collaboration, and energy systems operations. He has received awards including the IEEE Power & Energy Society Best Paper Award (2025) and was a finalist for the INFORMS Wagner Prize (2021). Education : PhD in Machine Learning, Georgia Tech (2017–2022) B.S./M.S. in Computer Science, BUPT (2010–2017) His research emphasizes spatio-temporal data analysis , decision making under uncertainty , and applications to energy systems, healthcare, and public policy. Notable projects include optimizing police zone design (Wagner Prize finalist) and enhancing grid resilience through robust optimization. He actively collaborates with institutions like Argonne National Laboratory and NSF-funded projects. Awards : Best Paper Award, IEEE Power & Energy Society (2025) Gen-AI Fellows (2024) Finalist, INFORMS Wagner Prize (2021) Advising & Grants : Advises PhD students Zekai Fan, Wenbin Zhou, and others Recipient of Block Center Seed Grant (2024), NSF funding (2024) His work spans energy resilience, public policy optimization, and causal inference in social systems. He co-leads the INFORMS Data Mining Society and reviews for top journals like Operations Research and Management Science.
