Summary
Overview
Work History
Education
Skills
Inventions
Websites
Timeline
Generic

Sanjay Kumar Singh

Bengaluru

Summary

Experienced and passionate AI/ML leader with over 20 years of experience in creating intelligent systems. I specialize in building AI and ML products and platforms, using the latest advancements in Generative AI and Agentic AI to create seamless user experiences. My expertise spans search, ranking, and natural language understanding (NLU), with a deep-seated love for applied and computational mathematics.

Overview

21
21
years of professional experience

Work History

Sr. Director AI ML Engineering

Optum
Bengaluru
04.2022 - Current
  • Prior Authorization AI/ML Platform: As a leader, I drove the creation of an AI/ML platform to automate prior authorization. I defined our technical roadmap and led a team to build a comprehensive solution using deep learning and NLP models to read clinical data. We also developed predictive models and a triage engine to approve low-risk cases automatically, which freed up our specialists. The platform used an Agentic AI framework to manage the entire process, from start to finish. This initiative cut manual effort by 75% and saved an estimated 70% in costs, while also improving patient experience and supporting our regulatory compliance.
  • Find and Price Care Engine: Led the strategic design and implementation of an AI/ML platform to transform the way members find and price care, directly improving member satisfaction and operational efficiency. My team and I tackled the core issue of provider search inaccuracy and billing surprises by developing an integrated platform. We engineered a semantic search and ranking engine that consolidated fragmented provider data, delivering highly accurate and relevant results to member queries. Concurrently, we built and deployed predictive models to offer personalized cost estimates, giving members the crucial information needed to make informed healthcare decisions and avoid financial surprises. This initiative not only empowered our members with unprecedented transparency but also directly reduced the operational burden on call centers, streamlining the member experience and driving significant improvements in overall satisfaction. This initiative reduced agent calls for appointments and billing issues by 85% in six months, directly improving member satisfaction and reducing operational costs.
  • Enterprise data interoperability engine: Led the design and implementation of a next-generation AI/ML platform to solve the company's most critical data interoperability challenges. We addressed the persistent problem of data fragmentation from various clinical and operational systems, which created massive ETL overhead and degraded core business functions. We developed a powerful, layered AI/ML system that functions as a single source of truth for data transformation. We built an Agentic AI framework with a suite of specialized models—from NLP for semantic enrichment to computer vision for handling unstructured documents—that could ingest and process data from any source. The system's intelligence was further refined using reinforcement learning with Direct Preference Optimization to ensure data integrity and seamless canonical transformations. This initiative not only automated complex ETL processes but also fundamentally improved data quality across the organization. The result was a 90% reduction in ETL operations costs for key departments, a significant decrease in errors that led to better customer service, and the seamless creation of longitudinal health records for research. This platform became the backbone for improving data-driven decisions and operational efficiency company-wide.
  • Conversational and Knowledge Management Platform: Led the development of an AI-powered conversational and knowledge management platform to automate high-volume, routine queries for providers, payers, auditors, and internal operations. The goal was to improve self-service, reduce agent workload, and provide proactive support. Architected and implemented a multi-layered platform leveraging advanced AI/ML. The solution featured NLU Models to accurately identify user intent, extract entities, and track sentiment and context across conversations. Knowledge Graphs to connect disparate data sources (members, health plans, medical history) and provide a comprehensive view for generating responses. Retrieval and Re-ranking Modules to efficiently source and present optimal information. Specialized models to assist Large Language Models (LLMs) with dynamic prompting, including triage models, medical coding, and cost optimization.
  • Strategic AI/ML Leader focused on Intellectual Property and Responsible Innovation. Drove the creation of intellectual property as a member of the Patent Review Board and ensured ethical, responsible development as part of the Board of AI Review and Responsible Use. Actively supported and prepared teams to present their cutting-edge research at top-tier academic conferences, including NeurIPS.

Director Data Science

Zycus
Bengaluru
05.2020 - 04.2022
  • Led the conception, architecture, upgradation and deployment of "Merlin," as a a company-wide AI/ML platform that became a key driver of product modernization at Zycus.
  • As a leader, I built and managed a high-performing team of AI scientists and engineers, fostering a culture of rapid innovation. Our work on the Merlin platform directly impacted Zycus's core products by Empowering executive decision-making: We deployed models for contract summarization and e-signature, giving COO, CPO, and CEO-level executives critical insights into legal and financial risks. Enhancing operational efficiency: Our models for search, category normalization, and product cataloging processed millions of items, significantly improving the user experience and system performance. Enabling data-driven strategy: We created sophisticated models for demand forecasting, negotiation, and supplier risk management that enabled more strategic sourcing decisions. Through this initiative, we not only modernized our product portfolio but also positioned Zycus as a leader in AI-driven enterprise solutions within the procurement business.

Lead Data Scientist

Optum
Bengaluru
05.2015 - 05.2020
  • Developed and managed an in-house Conversational AI platform to deploy high-performance chatbot applications for consumers and agents. The platform was designed to automate support activities related to claims, benefits, and eligibility, directly addressing a critical cost driver. I designed and implemented the core AI models, including: Entity Recognition: To accurately identify and extract key information from user queries. Context Recognition & Dialogue Management: To maintain conversational flow and handle complex user interactions. This project resulted in a significant reduction in support costs by lowering infrastructure, development, and maintenance overhead.
  • Claims Liability Prediction System: Developed Claims Liability Prediction System to automate and accelerate claims processing for millions of transactions per month. System was designed to predicts patient and payer liability, System performd with more than 87% accuracy. By replacing existing manual and rule-based workflows, this solution help cut down on the time required to validate claims, leading to an improvement in efficiency and processing accuracy. This initiative resulted in time and cost savings for the organization.
  • I took initiative to build a triage system for nurses with an objective to overhaul the existing manual and rule-based triage system. Solution was designed to analyze historical conversation data, official guidelines, and training documents to create an intelligent platform. This system not only streamlined the health triage process but also provided a more accurate and efficient service. The implementation resulted in a dramatic reduction in average call time, from 10 minutes down to just 2 seconds, which is a 30X improvement with an 85% accuracy rate. This initiative delivered significant operational savings and improved member satisfaction by optimizing the entire call experience.
  • Led the development and implementation of an AI/ML system to combat claims fraud, waste, and abuse. The platform achieved over 80% accuracy in identifying fraudulent transactions and included a continuous learning mechanism to adapt to evolving processes, resulting in a sustainable reduction in fraudulent activity.
  • Blockchain Innovation: Architected a new design pattern for blockchain applications that enabled complex, off-chain computations without compromising security. This solution was later adopted for enterprise-wide SMART on FHIRimplementations, demonstrating expertise in secure, scalable blockchain architecture.
  • AI/ML Application Development: Developed and deployed Optum's foundational deep learning chatbot for the AiDAplatform. Automated key business functions by creating and implementing machine learning models for IT, Networking, and HR systems.
  • System Development & Integration: Modified a Multichain platform to build a custom blockchain-based EHR system, later integrating it with providers, labs, and pharmacies. Additionally, developed a Predictive Monitoring System for real-time application health and performance.

Lead Engineer

Altisource
Bengaluru
01.2014 - 05.2015
  • In this role, I was responsible for the entire lifecycle of the Bidding Engine, Search, and Recommendation Engine, from initial requirements gathering and architectural design to hands-on development and deployment. My leadership focused on empowering the team to build a robust, high-performance platform that drove significant improvements in the bidding experience.

Manager

Accenture
Bengaluru
09.2004 - 01.2013
  • For five years (March 2008 - January 2013), I served as a dedicated on-site analytics consultant at StateFarm's headquarters in Bloomington, Illinois. In this role, I led analytics efforts to significantly boost sales of auto, home, and rental insurance. I was responsible for key projects including customer profiling, log analysis, and developing segmentation and recommendation models. The insights I provided to the sales and marketing teams were directly tied to enhancing sales strategies and improving overall performance against key performance indicators (KPIs).
  • Analytics Consultant, CUNA Mutual Group (March 2006 – March 2008) As an on-site consultant at CUNA Mutual's headquarters, I partnered with their marketing communications team to develop an intelligent analytics solution. I designed and built a software platform that precisely segmented 1.5 million customer records, enabling a more targeted and effective marketing strategy. Over two years, I was responsible for the full development and maintenance lifecycle of this product, which consistently performed with over 95% accuracy and a harmonic mean of over 91%, significantly improving the company's marketing efforts.
  • As a Software Consultant at TSYSAcquiring Solutions, I played a key role in the development of critical software systems at their Phoenix head office. I focused primarily on enhancing and maintaining the Account Clearance and Settlement systems, ensuring smooth and accurate financial transactions. Additionally, I contributed to the Merchant Management Systems, broadening my impact across the company's core business operations.

Software Engineer

AnMSoft Technologies
Mumbai
05.2004 - 09.2004
  • Product Development

Education

Advanced Program in Data Science - Computational And Applied Mathematics

Indian Institute of Management Calcutta
Joka West Bengal
03-2019

Bachelor of Engineering - Electronics And Communications Engineering

Visvesvaraya Technological University
Belgaum Karnataka
08-2002

Skills

  • TensorFlow, PyTorch, LSTM, CNN, RNN, GAN, GNN, Encoder, Decoder, Reinforcement Learning, Dynamic Memory Network
  • Random Forest, Information Gain, Entropy, IREP, RIPPER, KMeans, KNN, Apriori, ID3, CRF, HMM, Regression
  • Blockchain, Etherium, Quorum, Multichain
  • Python, Java, Scala, R, Neo4J
  • Search, Recommendation, Retrieval, Ranking, Graph, RAG
  • GenAI, AI Agents, Agentic AI, MCP
  • LoRA, QLoRA, Quantization, Distillation
  • NLP, NLU, NLG, Computer Vision
  • Responsible and Explainable AI
  • Product design, Architecture, Cloud platforms (Azure, AWS, GCP)
  • Team building, Communication, Program management, Healthcare domain knowledge

Inventions

  • Machine learning techniques for disambiguating unstructured data fields for mapping to data tables
  • Contextual document summarization with semantic intelligence
  • Hierarchical delimiter identification for parsing raw data
  • Canonical transformation using a machine learning language model
  • Document indexing, searching, and ranking with semantic intelligence
  • Systems and methods for contextual ranking of search results
  • Customization of forecasting solutions
  • System and method for classification of spend data
  • Nurseline Interface graphical interface for triage

Timeline

Sr. Director AI ML Engineering

Optum
04.2022 - Current

Director Data Science

Zycus
05.2020 - 04.2022

Lead Data Scientist

Optum
05.2015 - 05.2020

Lead Engineer

Altisource
01.2014 - 05.2015

Manager

Accenture
09.2004 - 01.2013

Software Engineer

AnMSoft Technologies
05.2004 - 09.2004

Advanced Program in Data Science - Computational And Applied Mathematics

Indian Institute of Management Calcutta

Bachelor of Engineering - Electronics And Communications Engineering

Visvesvaraya Technological University
Sanjay Kumar Singh