Samsudhin H
Chennai
Summary
Astute Staff Data Scientist with 14+ years of experience in data-driven and technology-focused approach. Communicates clearly with stakeholders and builds consensus around well-founded models. Talented in writing applications and reformulating models.
Overview
18
18
years of professional experience
1
1
Certification
Work History
Staff Data Scientist
ResMed India private ltd
02.2019 - Current
- Collaborated with stakeholders to devise quarterly roadmaps, prioritizing high-impact Generative AI initiatives and traditional ML projects based on effort and test coordination.
- Guided and supported junior data scientists in mastering modern AI techniques, including LLM fine-tuning, prompt engineering, and core data mining methodologies.
- Analyzed and evaluated data from hundreds of databases, leveraging Retrieval-Augmented Generation (RAG) and LLM-driven extraction to enhance product development and optimize business strategies.
- Utilized statistical techniques alongside Generative AI models to analyze complex datasets and automate the extraction of key insights.
- Employed mathematical techniques and advanced transformer architectures to formulate innovative engineering and scientific solutions.
- Discovered narratives conveyed by data, utilizing Natural Language Generation (NLG) and LLMs to clearly present automated, actionable insights to clients and business managers.
- Designed and architected the end-to-end lifecycle of deep learning models, integrating large language models with knowledge graphs to develop robust, context-aware scientific solutions.
- Engineered sophisticated algorithms by combining comprehensive statistical analysis, predictive data modeling, and foundational generative frameworks.
- Engineered and implemented hybrid predictive and generative AI models to inform strategic business decisions and automate complex workflows.
- Collaborated with stakeholders to formulate AI-focused quarterly roadmaps, evaluating the technical feasibility and business impact of deploying GenAI solutions into existing architectures.
Module Lead - Data Science
Cognizant Technology Solutions
12.2015 - 02.2019
- Spearheaded the design and product architecture of the Interactive Analytics Workbench (Data Science Workbench), establishing it as the core analytical engine within the BigDecisions Platform.
- Directed and mentored a team of data scientists, driving the development and deployment of sophisticated analytical algorithms to solve complex, real-world enterprise use cases.
- Architected and delivered StatisticsUtil, a foundational Java-based machine learning library that served as the technical backbone for the workbench's analytical capabilities.
- Led the engineering of SparkAnalytics, a highly scalable Scala machine learning library designed to power large-scale big data analytics across the platform.
- Orchestrated the development of a polyglot Analytical Pipeline, empowering users to seamlessly chain and execute multi-language analytical components as scalable batch operations.
- Drove the implementation of a comprehensive Data Exploration module, delivering advanced statistical profiling capabilities for end-users to evaluate data sets effectively.
- Championed the platform's MLOps infrastructure by designing the Model Management component, streamlining the end-to-end lifecycle for deploying models, publishing web services, and automating re-tuning and recalibration processes.
Associate Engineer
Cognizant Technology Solutions
08.2013 - 12.2015
- Spearheaded the foundational design and architecture of the nXg Device Analytics platform as a core data science team member.
- Engineered a dual-model Device Failure Prediction system, leveraging Accelerated Failure Time (AFT) for precise time-to-event forecasting and Logistic Regression for failure probability estimation.
- Applied collaborative filtering techniques to isolate, analyze, and predict the root causes of complex device failures.
- Developed prognostic statistical models to accurately estimate the remaining useful life and overall lifespan of deployed devices.
- Scaled predictive capabilities for big data environments by migrating legacy R-based analytical solutions into high-performance Spark Scala applications.
- Architected a near real-time predictive monitoring pipeline by developing a custom extension that seamlessly integrated R-based statistical models directly with QlikView.
- Designed interactive QlikView dashboards and built custom visualization extensions to dynamically map and analyze Root Cause networks for systemic device failures.
Software Engineer
Pigstick Software Solutions
07.2008 - 06.2010
- Designed and developed the core Java backend for a comprehensive Student Mark Sheet Management module, automating grading calculations and report generation.
- Engineered and optimized relational database schemas, writing efficient SQL queries to securely store, manage, and retrieve academic records.
- Conducted rigorous unit testing and debugging across the software development lifecycle, significantly reducing defects and ensuring high data accuracy.
- Collaborated with cross-functional teams and senior engineers to translate business requirements into clear technical specifications and scalable code.
- Assisted in the integration of backend Java logic with user-facing interfaces, streamlining the data-entry process for academic administrators and staff.
Education
Ph.D. - Generative AI And Knowledge Graphs (Pursuing)
BSA Crescent Institute of Science And Technology
Chennai, India08-2030
Master of Computer Applications - Computer Science And Applications
PSG College of Arts & Science
Coimbatore05.2013
Bachelor of Computer Applications - Computer Science And Applications
Texcity Arts & Science College
Coimbatore05.2008
Skills
- Generative AI
- Transformers & Large Language Models (LLMs)
- Knowledge Graphs & Graph Theory
- Natural Language Processing (NLP)
- Machine Learning & Deep Learning
- ML End-to-End Architecture
- MLOps & Model Management
- Cloud Platforms (AWS, Azure)
- Predictive Analytics & Statistical Analysis
- Apache Spark & Big Data
- RDBMS, NoSQL, Document Storage
- Predictive Analytics & Statistical Analysis
Certification
Machine Learning by Andrew NG, Coursera
Big Data Hadoop Foundations, IBM
Mathematical Foundation For Machine Learning and AI, Udemy
Deep Learning A-Z: Hands-On Artificial Neural Networks, Udemy
Blockchain and Deep Learning: Future of AI, Udemy
Neural Networks and Deep Learning, DeepLearning.AI
Structuring Machine Learning Projects, DeepLearning.AI
Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization, DeepLearning.AI
Timeline
Staff Data Scientist
ResMed India private ltd
02.2019 - Current
Module Lead - Data Science
Cognizant Technology Solutions
12.2015 - 02.2019
Associate Engineer
Cognizant Technology Solutions
08.2013 - 12.2015
Software Engineer
Pigstick Software Solutions
07.2008 - 06.2010
Bachelor of Computer Applications - Computer Science And Applications
Texcity Arts & Science College
Ph.D. - Generative AI And Knowledge Graphs (Pursuing)
BSA Crescent Institute of Science And Technology
Master of Computer Applications - Computer Science And Applications
PSG College of Arts & Science
Projects
Name: Intelligent Document Automation (IDA)
- Architected and deployed an end-to-end Intelligent Document Automation platform on AWS (EKS, Lambda, Bedrock), successfully processing 50,000+ medical documents monthly across 6+ live deployments with 92–95% extraction accuracy and ~1-minute latency.
- Spearheaded the AI/ML pipeline, integrating Large Language Models (LLMs), Named Entity Recognition (NER), Classification, and Contextual Reasoning to automate PHI extraction and determine insurance eligibility from unstructured clinical documents.
- Eliminated manual, paper-heavy workflows across major enterprise products (Brightree, MatrixCare, CitusHealth) by expanding the platform to support multilingual processing (English, Korean, German) for use cases including risk classification, care plan analysis, and discharge summarization.
- Engineered a robust, HIPAA-compliant security architecture ensuring full PHI protection with AES-256 encryption, 6-year immutable audit logs, AWS BAA coverage, and zero critical vulnerabilities across SAST, SCA, and container scans.
- Accelerated production release cadence by 3x (transitioning from quarterly to monthly cycles) and reduced annual refactoring by 67% by strategically integrating AI-assisted development tools (GitHub Copilot, Augment AI) to improve overall code quality.
Name: Payment Engagement Score (PES)
The Payment Engagement Score (PES) is a metric designed to measure the level of interaction between consumers and providers regarding payments.
Name: Clinical Advanced Insights
Site: https://www.matrixcare.com/clinicaladvancedinsights/
This product is the first AI/ML based product build in MatrixCare, it started with 5 members and I am one of the founding members of the team. It has produced $4.4 million dollars as revenue in short span of time. It has 3 solutions as listed below,
Fall Risk Analysis:
It’s an ensemble model which combines weighted formulae based on clinical experts input and Neural Network. Goal of the project is to predict fall risk among senior citizens and reasons for the fall in skilled nursing facilities (SNF). It helped clients to identify early fall risk of the residents and prevent it by adding appropriate care plans.
Change In Condition Analysis:
It’s a formula based mathematical model that can able to identify changes in resident’s clinical condition in real time that triggers clients to optimize their staff and allocate care units as per resident’s clinical condition.
COVID-19 Severity Risk Analysis:
It’s an ensemble model which combines weighted formulae based on clinical experts input and Logistic Regression. This model is created in short span of time in the urgency of COVID-19. It helped the clients to identify and categories residents based on their risk severity to organize COVID Vaccines and respective treatments.
Name: BigDecisions
Site: https://www.youtube.com/watch?v=y7Ue_7relRg
BigDecisions is Cognizant in build data platform that supports end to end data solutions.
Name: nXg Device Analytics
Most digital information today is captured with the aid of devices, hence device downtimes can have adverse impact on operations and thereby profits. Where devices are installed at the customer site, it is imperative that the service is not disrupted by device failure. nXg uses advanced analytics to accurately predict device failures and generate a holistic picture of device health and to improve overall business performance and customer satisfaction.