Digital Einstein
An embodied conversational AI agent bringing Albert Einstein back to life through multimodal interaction — combining speech synthesis, facial animation, and cognitive reasoning. …
Professional Summary
I am a Senior Researcher & Lecturer at the Computer Graphics Laboratory of ETH Zurich, and a Research Consultant at Disney Research. I am leading the Digital Character AI projects at CGL. My research interests include conversational digital characters, affective computing, human-computer interaction, and applied machine learning.
My vision is to create intelligent digital humans that can naturally communicate, understand, and support people across domains such as education and mental health. My research focuses on multimodal artificial intelligence for interactive digital humans, developing models that combine large language models, affective computing, and data-driven animation to create embodied conversational agents endowed with autonomous agency, consistent values, and beliefs.
My work bridges machine learning, human–computer interaction, and computer graphics to enable AI systems such as Digital Einstein and interactive patient avatars for psychotherapy training and health education.
Education
Dr. sc. (PhD) in Computer Science
ETH Zurich
Master's Degree in Computer Science
ETH Zurich
Bachelor's Degree in Computer Science
ETH Zurich
Interests
Embodied Conversational Agents
Multimodal AI and Affective Computing
Human–AI Interaction
Data-Driven Animation Synthesis
LLMs & Generative AI for Dialogue
AI for Mental Health
My Research
My research addresses a fundamental challenge at the intersection of deep learning and human-centric AI: how can we build machine learning systems that not only perform with high accuracy but also meaningfully understand, adapt to, and collaborate with humans?
I work across three interconnected areas: multimodal deep learning for affective and behavioral understanding, leveraging smartphone sensors, egocentric vision, and physiological signals; agentic AI and cognitive architectures for conversational agents, enabling consistent personality, beliefs, and emotional states; and scalable generative models for embodied expression, developing efficient transformer pipelines for text-driven and speech-driven facial animation.
A flagship application is Digital Einstein, an embodied conversational agent that has reached thousands of visitors globally at events including SIGGRAPH, GITEX Global, and the World Economic Forum. My system also powers collaborative applied research in digital health — including a Virtual Psychotherapist for mental health support and RehaBot for patient health education.
Please reach out to collaborate.
Research Projects
Affective Computing & Emotion Recognition
Multimodal deep learning systems for affective state prediction from smartphone sensors, biometric data, egocentric vision, and physiological signals — enabling real-world …
Personality & Cognitive Architectures for Conversational Agents
Deep learning-based cognitive frameworks that enable conversational agents to maintain consistent personality, beliefs, values, and emotional states across extended interactions — …
Facial Animation Synthesis
Efficient transformer-based pipelines for text-driven and speech-driven 3D facial animation — achieving real-time performance through phoneme-level speech modeling and contrastive …
Dialog Act Classification
Multimodal dialog act classification for conversations with digital characters — interpreting the communicative intent of user utterances to drive appropriate agent responses in …
Virtual Psychotherapist
Embodied conversational agents for mental health support, developed in partnership with Prof. Andrew Gloster (University of Lucerne). The project comprises a patient-facing …
RehaBot
An embodied AI avatar representing a medical professional, capable of conducting patient interactions, health assessments, and health education in rehabilitation and home-care …
MIND — Cognitive Health Monitoring
A mobile AI platform for early detection of cognitive impairment in aging populations, integrating Large Geospatial Models for movement analysis, LLMs for conversational cognitive …
Funding
My research is supported by competitive grants from national and international funding agencies. Current projects include work on cognitive health monitoring funded by the National Research Foundation Singapore (NRF CREATE – Future Health Technologies 2), animation synthesis funded by an SNF Ambizione grant, and applied digital health projects in collaboration with Swiss universities and hospitals. Past funding includes an ETH Grant and project partnerships with the Hasler Foundation and the University of Basel.
Selected Talks
- Jan 2026 — AI Literacy and the Future of Work, Panel discussion, World Economic Forum, Davos
- Nov 2025 — Beyond Avatars, Panel discussion, Berlin Science Week, Berlin
- Oct 2025 — A Platform for Interactive AI Character Experiences, Invited talk, GITEX Global, Dubai
- Aug 2025 — Digital Einstein at SIGGRAPH, Vancouver
- Oct 2024 — The Creation of Believable Digital Characters, Keynote, GITEX Global, Dubai
People
I supervise a team of PhD students and research assistants. I currently supervise three PhD students, and have successfully supervised one PhD to completion. Current PhD projects span user interface optimisation, believable narrative agents, and data-driven animation synthesis for conversational characters. I have supervised 35 Bachelor's, 13 Master's, and 7 semester thesis projects, as well as four student research assistants working on the Digital Einstein project.
Featured Publications
PhonemeNet: A Transformer Pipeline for Text-Driven Facial Animation
A transformer pipeline for text-driven facial animation exploiting phoneme-level speech structure, achieving real-time performance and best-in-class lip synchronization accuracy. …
p.-witzig
egoEMOTION: Egocentric Vision and Physiological Signals for Emotion and Personality Recognition in Real-World Tasks
A new multimodal dataset and architecture combining egocentric vision and physiological signals for in-the-wild emotion and personality recognition, presented at NeurIPS 2025 …
m.-jammot
A Platform for Interactive AI Character Experiences
A full-pipeline platform for interactive AI character experiences, demonstrated through Digital Einstein and deployed at scientific conferences, technology events, and public …
The Personality Dimensions GPT-3 Expresses During Human-Chatbot Interactions
A large-scale empirical characterization of the personality dimensions GPT-3 expresses during human-chatbot interaction, using Big Five psychometrics. Published in ACM IMWUT 2024.
n.-kovacevic
Affective State Prediction from Smartphone Touch and Sensor Data in the Wild
Multimodal affective state prediction from smartphone touch and sensor data in naturalistic conditions, using deep learning fusion. CHI 2022.
All Publications
(2026).
Talking to Your Data: Exploring Embodied Conversation as an Interface for Personal Health Reflection.
In IUI Workshops 2026.
(2025).
PhonemeNet: A Transformer Pipeline for Text-Driven Facial Animation.
In MIG 2025.
(2025).
egoEMOTION: Egocentric Vision and Physiological Signals for Emotion and Personality Recognition in Real-World Tasks.
In NeurIPS 2025.
(2025).
Steering Narrative Agents through a Dynamic Cognitive Framework for Guided Emergent Storytelling.
In AAAI AIIDE 2025.
(2025).
A Joint Personality-Emotion Framework for Personality-Consistent Conversational Agents.
In IVA 2025.
Teaching
| Course | Role | Institution | Years |
|---|---|---|---|
| Artificial Intelligence for Digital Characters | Course creator & lecturer | ETH Zurich | 2024 – present |
| Seminar on Digital Humans | Course creator & lecturer | ETH Zurich | 2022 – present |