Shantanu Acharya
Research Scientist at NVIDIA
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolShantanu Acharya is a Research Scientist at NVIDIA with a decade of experience building and optimizing deep learning systems, currently focused on long-context capabilities of large language models. He holds an MS in Computer Science from NYU Courant and has blended academic research in computer vision and NLP with applied work in speech and multimodal models across NVIDIA and NYU. His hands-on contributions to the widely used NVIDIA NeMo framework include performance-minded data preprocessing, Conformer encoder adaptations, and architecture/configuration improvements that bridge research and production. Comfortable across audio, vision, and language domains, he has a track record of shipping reproducible code, improving tooling, and squeezing efficiency from model pipelines. Based in New York, he brings both rigorous research instincts and pragmatic engineering discipline to scale large-model capabilities.
10 years of coding experience
4 years of employment as a software developer
Bachelor of Technology - BTech Computer Science and Engineering, Bachelor of Technology - BTech Computer Science and Engineering at National Institute of Technology MizoramMaster of Science - MS Computer Science, Master of Science - MS Computer Science at New York University
English, Bangla, Hindi
Stackoverflow
87reputation
5kreached
5answers
2questions
Github Skills (20)
pytorch10
python10
machine-learning10
automatic-speech-recognition10
data-preprocessing10
deep-learning10
asr10
neural-network10
speech-recognition10
python-multiprocessing9
large-language-models9
speech-synthesis9
multiprocessing9
multi-process9
generative-ai8
Programming languages (7)
Github contributions (5)
NVIDIA/NeMo
Jun 2022 - Oct 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:22 reviews, 8 commits, 17 PRs in 4 months
Contributions summary:Shantanu primarily contributes to the NVIDIA NeMo framework, focusing on improving the speech commands dataset processing script by incorporating multiprocessing support for faster data preprocessing. They address code quality by fixing styling issues, adding docstrings, and correcting bugs related to silence set construction. Furthermore, the user works on adapting the Conformer encoder, which includes adjusting input dimensions, auto-switching adapters, and adding support for specifying audio dropout, showcasing an understanding of model architecture and configuration for improved performance.
asrspeech-recognitionnatural-language-processingttsspeaker-diarization
shan18/TensorNet
Mar 2020 - Nov 2021
A high-level deep learning library built on top of PyTorch.
Contributions:189 commits, 40 PRs, 115 pushes in 1 year 8 months
pytorchcaffe2efficientnetdeep-learninghigh-level
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.