Subhankar Ghosh
Senior Applied Research Scientist at NVIDIA
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolSubhankar Ghosh is a Senior Applied Research Scientist at NVIDIA with nine years of experience applying mathematical rigor to machine learning and deep learning, particularly in generative speech and LLMs within the NeMo framework. He blends a statistics MS from UIUC with strong software engineering chops honed at Microsoft and Google to move research into production-grade systems. Notable contributions include advancing multi-speaker and energy-conditioned FastPitch Text-to-Speech capabilities in NVIDIA/NeMo, reflecting deep expertise in Speech AI and multimodal generative models. Based in New York, he pairs probabilistic modeling and stylometry research background with practical engineering to solve real-world NLP and speech problems. Colleagues value his ability to translate theoretical insights into scalable tutorials and tooling that accelerate developer adoption.
9 years of coding experience
4 years of employment as a software developer
Master of Science - MS Statistics with Analytics Concentration, Master of Science - MS Statistics with Analytics Concentration at University of Illinois Urbana-ChampaignBachelor of Technology (B.Tech.) Computer Science and Engineering, Bachelor of Technology (B.Tech.) Computer Science and Engineering at National Institute of Technology RourkelaSt Thomas Boys School
English, Bengali, Hindi
Github Skills (8)
pytorch10
machine-learning10
text-to-speech10
large-language-models10
generative-ai10
python9
neural-network9
deep-learning9
Programming languages (3)
Github contributions (5)
NVIDIA/NeMo
Feb 2022 - Nov 2022
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
Role in this project:
ML Engineer
ML EngineerContributions:158 reviews, 49 commits, 48 PRs in 9 months
Contributions summary:Subhankar's primary contributions revolve around the Text-to-Speech (TTS) domain within the NeMo framework. They have added and refined FastPitch training tutorials, demonstrating a focus on large language models and generative AI. Furthermore, the user has made changes to support multi-speaker FastPitch models and speaker interpolation, enhancing the model's capabilities. Key additions include energy conditioning and speaker embedding conditioning.
asrspeech-recognitionnatural-language-processingttsspeaker-diarization
subhankar-ghosh/subhankar-ghosh.github.io
Sep 2017 - Mar 2022
Contributions:203 pushes, 1 branch in 4 years 6 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.