John Murzaku

John Murzaku

Applied Scientist @ Adobe

PhD in Computer Science from Stony Brook. At Adobe, I build AI agents focused on disambiguation, context, and memory. My research bridges NLP with cognitive science.

01 Curriculum Vitae 02 jam@lissus.com 03 Google Scholar 04 GitHub 05 LinkedIn

Research

12 publications · click to read

Belief & Theory of Mind
Zero-Shot Belief TrackingCODI 2025 Views Are My Own, but Also YoursACL Findings 2024 BeLeaf: Belief as Tree GenerationNAACL Demo 2024 Generative Event FactualityACL Findings 2023 Re-Examining FactbankCOLING 2022
Multimodal
OmniVox: Emotion RecognitionarXiv 2025 Synthetic Audio for Cognitive TasksNAACL Findings 2025 Multimodal Belief PredictionInterspeech 2024
Psycholinguistics
Projecting Knowledge & Common GroundToM4AI @ AAAI 2025 Recognizing Hedges in NarrativesSIGDIAL 2024
Applied AI
ECLAIR: Enhanced ClarificationIAAI 2025 ECLAIR DemoAAAI Demo 2025
more below

Updates

Jan 2026
Now working on disambiguation, context, and memory for AI agents at Adobe.
May-Dec 2025
Defended PhD! Joined Adobe in San Jose, working on planning and agentic systems for the Agent Orchestrator team.
Jan 2025
ECLAIR accepted to AAAI Demo + IAAI. Synthetic audio paper at NAACL Findings.
2024
Adobe internship → 2 papers + patent. Publications at SIGDIAL, ACL Findings, Interspeech, NAACL Demo.

All Publications

Zero-Shot Belief: A Hard Problem for LLMsBelief
CODI 2025
J. Murzaku, O. Rambow
OmniVox: Zero-Shot Emotion Recognition with Omni-LLMsMultimodal
arXiv 2025
J. Murzaku, O. Rambow
Projecting Knowledge and Common Ground from Characters' UtterancesPsycholing
ToM4AI @ AAAI 2025
A. Soubki, A. Paige, J. Murzaku, O. Rambow, S. Brennan
ECLAIR: Enhanced Clarification for Interactive Responses (Demo)Adobe
AAAI 2025 Demo
J. Murzaku, Z. Liu, V. Muppala, M. Tanjim, X. Chen, Y. Li
ECLAIR: Enhanced Clarification for Interactive ResponsesAdobe
IAAI 2025 · 15.7% acceptance
J. Murzaku, Z. Liu, M. Tanjim, V. Muppala, X. Chen, Y. Li
Synthetic Audio Helps for Cognitive State TasksMultimodal
Findings of NAACL 2025
A. Soubki*, J. Murzaku*, P. Zeng, O. Rambow
Training LLMs to Recognize Hedges in Spontaneous NarrativesPsycholing
SIGDIAL 2024
A. Paige*, A. Soubki*, J. Murzaku*, O. Rambow, S. Brennan
Views Are My Own, but Also Yours: Benchmarking Theory of MindBelief
Findings of ACL 2024
A. Soubki, J. Murzaku, A. Y. Jordehi, P. Zeng, M. Markowska, S.A. Mirroshandel, O. Rambow
Multimodal Belief PredictionMultimodal
Interspeech 2024
J. Murzaku*, A. Soubki*, O. Rambow
BeLeaf: Belief Prediction as Tree GenerationBelief
NAACL 2024 Demo
J. Murzaku, O. Rambow
Towards Generative Event Factuality PredictionBelief
Findings of ACL 2023
J. Murzaku, T. Osborne, A. Aviram, O. Rambow
Re-Examining Factbank: Predicting the Author's Presentation of FactualityBelief
COLING 2022
J. Murzaku, P. Zeng, M. Markowska, O. Rambow