Binesh Bannerjee
Associate Director of Artificial Intelligence and Machine Learning @Penguin Random House
Signup · Get unlimited contacts
WORK HISTORY
Associate Director of Artificial Intelligence and Machine Learning @Penguin Random House
New York, NY, US
Led the development of Simile, a multi-modal neural network using PyTorch and TensorFlow for book classification, enhancing metadata accuracy with diverse embeddings and binary codes. According to a internal Random House study, this work boosted sales by 3.8% through Simile’s outgoing tags for Amazon, optimizing book metadata with advanced machine learning algorithms.Created a 3D visualization for book clusters using three.js based on Simile’s embeddings, improving data presentation and decision-making for internal stakeholders.Developed a Python wheel for Simile, simplifying integration and boosting its adoption across the company, enhancing productivity.Unified diverse data points (users, authors, search terms) into a single space using machine learning, enhancing discoverability and relevance across the publishing ecosystem.Developed and deployed a GPT-powered chatbot, producing contextually relevant book descriptions to aid marketing efforts.Enhanced user interaction with Retrieval-Augmented Generation (RAG) and GPT models, improving conversational responses and search term matching with embeddings produced by LLM models from Huggingface (We experimented with Longformer, Bloomz among others).Combined Simile embeddings with GPT to create and label book clusters, advancing content management and recommendation systems.Initiated and led \'Python for Non-Programmers\' and evolved it into \'Processing for Programmers\', teaching diverse company departments from metadata to sales. I championed the philosophy of automating tedious job aspects, inspiring participants to embrace programming as a tool for enhancing efficiency and innovation. The program included engaging projects with microcontrollers like Microbit and Arduino, fostering technical skill development and creativity across the company.Mentored a team on AI projects including Stable Diffusion and neural focus analysis with GradCAM, uncovering insights into model performance.
EDUCATION
Queens College
BA, Computer Science
SKILLS
ABOUT BINESH BANNERJEE
With over 18 years of experience in data analytics and machine learning, I specialize in harnessing AI to revolutionize the publishing industry. As the Associate Director of Artificial Intelligence and Machine Learning at a leading publishing house, I spearheaded the development of Simile, an innovative multi-modal neural network. This system classifies books based on the text, cover art, flap copy, and tags, significantly enhancing metadata quality and discoverability.My expertise extends to designing and implementing advanced AI solutions using PyTorch, TensorFlow, numpy, and scikit-learn, creating a unified, scalable framework that incorporates diverse data points such as users, authors, and search terms. A notable achievement includes deploying Simile\'s technology to generate precise outgoing tags for Amazon, directly contributing to a 3.8% sales uplift by refining keyword metadata for our publications.I am passionate about cross-departmental collaboration, working closely with metadata, sales, and content teams to forge new paths in AI application, ensuring that technology not only meets but anticipates industry needs. My mission is to leverage AI not only to optimize operations but also to enhance reader satisfaction and drive business success through informed, data-driven strategies.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.