Navin Raj PrabhuN. R. Prabhu

Curriculum vitae

Researcher in artificial intelligence, speech processing and social signal processing, with experience in both academic and start-up teams.

Research interests

  • Generative modeling
  • Probabilistic and uncertainty modeling
  • Speech emotion recognition
  • Emotion-conditioned speech synthesis
  • Speech emotion conversion
  • Social signal processing
  • Group affect
  • Multimodal learning

Education

  • 2021 – 2026

    PhD in Computer Science

    University of Hamburg, Germany

    • Dissertation: Probabilistic and Generative Modeling for Emotion Recognition and Synthesis, defended February 2026
    • Supervisors: Prof. Dr.-Ing. Timo Gerkmann (Signal Processing) and Prof. Dr. Nale Lehmann-Willenbrock (Industrial and Organizational Psychology)
    • Probabilistic neural networks for uncertainty modeling in emotion recognition
    • Diffusion- and GAN-based generative models for emotion-conditioned speech synthesis
  • 2018 – 2020

    MSc in Computer Science, Data Science track

    Delft University of Technology, The Netherlands

    • Thesis: Conversation Quality: Modeling in Free-Standing Conversational Groups
    • Supervisors: Dr. Hayley Hung and Dr. Chirag Raman
    • Courses included Pattern Recognition, Deep Learning, Multivariate Data Analysis and Social Signal Processing
  • 2011 – 2015

    BTech in Computer Science

    SRM University, Chennai, India

    • Courses included Neural Networks, Pattern Recognition, Probability and Queuing Theory and Genetic Algorithms

Experience

  • Now

    Machine Learning Researcher

    ai-coustics, Berlin, Germany

    • Working on audio intelligence for voice AI at ai-coustics.
  • 2021 – 2025

    Doctoral researcher

    Signal Processing group, University of Hamburg

    • Interdisciplinary project with the Industrial and Organizational Psychology group on the recognition and synthesis of affect
  • 2020

    Research assistant

    Interactive Intelligence Lab, TU Delft

    • Conversational AI and social robots for social skills training
    • Adaptive robot behaviour based on a partner's behavioural cues, using FurhatSDK, OpenFace and Microsoft PSI
  • 2019

    Data science intern

    Momo Medical, Delft

    • Moved pose detection from heuristics to data-driven modeling with time-series neural networks
    • Designed the data collection pipeline from an exploratory analysis of data variance
  • 2015 – 2018

    Full-stack developer, founding employee

    Terraa Farm Fresh, Bangalore

    • Worked across iOS development, backend systems, deep learning and operations research for an online grocery platform
    • Built order allocation and subscription ledger systems, demand forecasting (TensorFlow, Google Cloud) and analytics dashboards (BigQuery)

Teaching and supervision

Teaching assistance at the University of Hamburg.

  • 2021 – 2025

    Supervised 7 bachelor's and 4 master's theses

  • Summer 2025

    Speech Signal Processing

  • Winter 2024/25

    Deep Learning for Multimodal Data Science; Deep Learning for Signal Processing

  • Summer 2024

    Machine Learning

  • Winter 2023/24

    Software Engineering

Publications

11 peer-reviewed papers and chapters, including IEEE Transactions on Affective Computing, ICASSP, Interspeech, ASRU and ACII.

See all publications