Anand Mohan
Applied Scientist at Amazon AGI
7+ years of applied research in large-scale Speech and Language AI. Leading continual learning and hot-fixing pipelines for Amazon's speech-to-speech LLMs, implementing fine-tuning, preference learning, and model alignment for production systems. Expert in parameter-efficient fine-tuning and responsible AI practices.
Applied Scientist - AGI
Amazon • Cambridge, UK
Leading Continual Learning and Hot-fixing for speech-to-speech LLMs. Driving model alignment, supervised fine-tuning, and preference learning for Alexa+ systems.
Applied Scientist - AGI
Amazon • Hyderabad, India
Built production ASR systems, developed federated learning frameworks, and led multilingual speech recognition initiatives.
Masters in Artificial Intelligence
Indian Institute of Science • Bangalore, India
Specialized in speech recognition and language modeling, published research on attention-based models.
Software Engineer - Site24x7
Zoho • Chennai, India
Full-stack development for performance monitoring platform, built AngularJS web clients and Java APIs.
Bachelors in Electronics & Communication
National Institute of Technology • Calicut, India
Foundation in signal processing and communication systems, elected student representative.
Publications
AMuSE: Attentive Multilingual Speech Encoding for Zero-Prior ASR
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) • 2025
Cross-silo Federated Training in the Cloud with Diversity Scaling and SSL
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) • 2021
Towards Relevance and Sequence Modeling in Language Recognition
IEEE Transactions on Audio, Speech and Language Processing • 2020
Attention based Hybrid I-vector BLSTM Model for Language Recognition
Annual Conference of the International Speech Communication Association (INTERSPEECH) • 2019
End-to-End Language Recognition Using Attention Based Hierarchical GRU Models
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) • 2019
The LEAP Speaker Recognition System for NIST SRE 2018 Challenge
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) • 2019
Technical Skills
Specializations
Post-Training (Supervised Fine-Tuning, Preference Learning), Parameter-Efficient Fine-Tuning (LoRA), Large Language Models, Distributed Training, Model Alignment, Production ML Systems, ASR, S2S, LID
Frameworks
PyTorch, PySpark, TensorFlow, Kaldi-ASR