Binesh Bannerjee
Associate Director Of Artificial Intelligence And Machine Learning
New York, New York, United States
Summary
👤
Senior🎓
Top SchoolBinesh Bannerjee is an AI and machine learning leader with over 18 years of experience applying advanced models to transform publishing operations and discoverability. As Associate Director of AI/ML at Penguin Random House he led development of Simile, a multi-modal neural network that blends text, cover art, metadata and user signals to improve tagging and drove a measurable 3.8% sales uplift via optimized Amazon metadata. He’s hands-on with PyTorch, TensorFlow and embeddings work, contributed practical improvements to the popular tensorboardX project, and built production-ready tooling (a Python wheel) and 3D visualizations to make model outputs actionable for business teams. A skilled translator between technical and non-technical stakeholders, he also created company-wide training programs that turned non-programmers into productive automation adopters. Binesh combines deep research-style experimentation (RAG, LLMs, Stable Diffusion) with pragmatic product delivery to boost both reader experience and commercial outcomes.
11 years of coding experience
25 years of employment as a software developer
CourseraBA Computer Science, BA Computer Science at Queens College
English
Stackoverflow
820reputation
62kreached
7answers
18questions
Github Skills (15)
tensorboard10
pytorch10
machine-learning10
python10
visualization8
numpy8
visualizations8
neural-network6
salt6
perl6
ant6
passwords6
password-encryption6
cryptography6
backpropagation6
Programming languages (10)
Github contributions (5)
lanpa/tensorboardX
Mar 2018 - Mar 2018
tensorboard for pytorch (and chainer, mxnet, numpy, ...)
Role in this project:
ML Engineer
ML EngineerContributions:6 commits, 1 PR, 3 comments in 12 days
Contributions summary:Binesh primarily contributed to the functionality of the `tensorboardx` library, which provides TensorBoard visualizations for PyTorch models. Their work included adding features such as handling multiple embeddings per epoch, demonstrating a practical use case. They also made code modifications by reorganizing the embedding directory structure and correcting the formatting of the code. They also made changes to the demonstration file and updated some of the code to be more concise.
pytorchmxnetnumpydeep-learningmachine-learning
bnsh/GPU-HW
Aug 2014 - Dec 2015
Contributions:121 commits, 13 pushes in 1 year 4 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.