Leading AI Innovation: From Foundational Research to Product Impact

Discover My Journey
Research and Product Balance Illustration

About Me

Photo

I am a Senior AI & Data Science Leader with over 25 years of experience, deeply focused on solving complex enterprise challenges through machine learning, large language models, and data mining. My expertise spans the design and deployment of scalable AI systems that drive automation, personalization, and decision-making across diverse sectors including digital marketing, infrastructure, and business process services.

My career journey uniquely bridges rigorous academic research with impactful product innovation. I've consistently translated foundational research breakthroughs into globally deployed solutions, recognized through influential publications, a portfolio of over 70 patents, and multiple industry and corporate awards.

I am passionate about aligning cutting-edge AI research with long-term business strategy to deliver measurable impact. My areas of expertise include Large Language Models (LLMs), Deep Learning, Natural Language Processing (NLP), Federated Learning, Model Compression, Scalable AI Deployment, Business Automation & Optimization, and AI Research Strategy.


Here's a selection of my most impactful projects, spanning both my academic research and product development career.

Research

RAG based LLM for Keyword Recommendation

Designed and deployed a keyword recommendation system powered by Large Language Models, enhancing targeting precision and improving campaign performance by 70%.

Product

Foundation Models for Customization

Implemented advertiser-level customized budget recommendations by training new foundation models based on transformer architectures, increasing adoption rates by 40%.

Research

Multimodal LLMs for Document Understanding

Deployed multi-modal transformer models for key-value pair (KVP) extraction, improving document extraction accuracy by 45% across procure-to-pay workflows for invoices and POs.

Product

OCR Optimization

Led efforts to enhance OCR performance on noisy documents, increasing text recognition accuracy by 73% through advanced data processing and DL model (CNN) enhancements.

Research

Distributed Training Benchmark

Surpassed global benchmark by training ResNet-50 on ImageNet-1K within 45 minutes using 256 GPUs, improving training efficiency by 25%.

Product

Conversational AI Deployment

Led integration of automated question-answering systems with Watson Assistant, improving customer engagement and satisfaction scores significantly compared to rule-based systems.


Get in Touch

I'm always open to discussing new research opportunities, product challenges, or collaborations. Feel free to connect!