Leading VNS Health's transformative generative AI initiative, I steer the innovative application of AI to revolutionize call center operations. This involves parsing and structuring vast amounts of unstructured audio data from members interactions, thereby enhancing care delivery and operational efficiencies. My role is pivotal in integrating cutting-edge AI with healthcare services, demonstrating significant advancements in member engagement and support.Key Contributions:• Generative AI Implementation: Spearheaded the deployment of generative AI to transcribe, parse, and structure call center audio calls, leveraging advanced open-source algorithms. This initiative marks a significant leap forward in utilizing AI to improve healthcare outcomes and customer service experiences.• ML Architecture Overhaul: Co-Led a collaborative effort to revamp our ML architecture, employing AWS SageMaker, Apache Airflow, and Feast feature store, ensuring a scalable, reliable, and efficient infrastructure.• Rapid ML Deployment: Introduced a CI/CD-based architecture for ML, facilitating swift model development and deployment, significantly reducing the productionization timeline.• Cross-Disciplinary Leadership: Coordinated with cross-functional teams to establish ML best practices, fostering an environment of innovation and excellence in AI application within healthcare.• In addition to leading generative AI initiatives, I am deeply involved in integrating state-of-the-art Language Model (LLM) techniques to further enhance our AI-driven solutions. Techniques: LoRA, QLoRA, AWQ, GPTQ, SGLang, DeepSpeed, vLLM, FP6-LLM, LangChain, Guidance, LMQL, RAG, LlamaIndex, PEFTTechnology Stack: Python, R, AWS SageMaker, Apache Airflow, Snowflake, Node.js, JavaScript, SQL.

• Build ML models for production use.• Perform in-depth exploratory data analysis to scope ML opportunities, identify potential issues in the data, and lay the groundwork for model design, training & evaluation strategies.• Suggest and implement new features to improve model performance and business logic.• Write and review production-quality PRs daily.• Translate client business problems into ML problems.• Build scalable machine learning models for claim time series forecasting and NLP text understanding.Technology Stack: Python, Tensorflow, PyTorch, Dagster, BigQuery, Vertex AI, GCP, Terraform, Spark, Kubernetes.

• Working hand-in-hand with our industry partner on the cutting edge of the intersection of databases, knowledge graphs, and artificial intelligence.• Contribute to the development of a product for solving real-world, large-scale problems using knowledge graph technology• Design and development of artificial intelligence and machine learning applications built upon the knowledge graph technology.Technology Stack: Python, Azure Databricks, PySpark, Palantir, REL, Julia, SQL.

• Applying natural language processing methods to clinical text to extract structured information.• Using the latest deep learning techniques to classify imaging studies.• Applying statistical models (with a focus on Bayesian methods) to assist researchers in analyzing missing, erroneous, or incomplete patient data.• Implement statistical and machine learning models, large-scale, cloud-based data processing pipelines, and off the shelf solutions for test and evaluation; interpret data to assess algorithm performance.• Develop novel ways to apply published machine learning models to imperfect clinical data including the development of training datasets.• Develop high-quality, secure code implementing models and algorithms as application programming interfaces or other service-oriented software implementations.• Manage and scale applications using container technology and cloud-hosted managed services.Technology Stack: Python, Apache Spark, BigQuery, Kubernetes, Argo, Dataflow, Apache Beam, Terraform, Docker, R

• Develop, build, test and deploy machine learning algorithms to support development of business processes for healthcare organization and subsidiary health plan provider, to improve business outcomes and quality of care.• Create and maintain framework for deploying machine learning algorithms using APIs.• Utilize resulting applications to implement, track and monitor predictive models used to guide business decisions.• Engineer computational solutions and develop algorithms and applications to meet the predictive needs of clinical and business units across the Visiting Nurse Service of New York.• Identify clusters of sub-populations of patients who may benefit from targeted care-management strategies: improves positive predicted value for patient outcomes based on sub-modeling for each cluster.• Ensure accuracy of deployed algorithms is monitored on an ongoing basis; alert management when algorithm performance declines, identify causes.• Ensure data quality throughout all stages of acquisition and processing, including sourcing, collection, ground truth generation, normalization & transformation.Technology Stack: R, SQL, Python, H2O, Random Forest, XgBoost, LightGBM, Regression, Clustering, KNN, GBM, PCA, SVD

• Design statistical models/programs using R/Python to successfully test hypotheses and answer targeted questions in healthcare outcomes research.• Design algorithms for the recommended actions/interventions that will best prevent adverse events and deterioration.• Cluster analysis for identifying sub-populations of complex patients who may benefit from targeted care management strategies and improvising positive predicted value for patient outcome by sub-modelling on each cluster.• Perform analysis on the targets including hospital readmission among Acute Myocardial Infarction (AMI) patients, IP visits, ER visits, MRSA among diabetes patients, congestive heart failure,Clostridium difficile (C. diff.), pressure injury,sepsis and fall-injury.• Writing complex SQL queries for data investigation and mapping to extract data foranalysis.• Build algorithms to reduce predictive analytics in driving population-level insightsand the expected ROI from predictive population health analytic solutions.• Report and visualize results of statistical analyses, in the form of graphs, charts, andtables using Tableau.𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 𝗦𝘁𝗮𝗰𝗸: 𝗥, 𝗦𝗤𝗟, 𝗣𝘆𝘁𝗵𝗼𝗻, 𝗔𝗽𝗮𝗰𝗵𝗲 𝗦𝗽𝗮𝗿𝗸, 𝗧𝗮𝗯𝗹𝗲𝗮𝘂, 𝗛𝟮𝗢, 𝗥𝗮𝗻𝗱𝗼𝗺 𝗙𝗼𝗿𝗲𝘀𝘁, 𝗫𝗚𝗕𝗼𝗼𝘀𝘁, 𝗛𝗗𝗕𝗦𝗖𝗔𝗡 𝗖𝗹𝘂𝘀𝘁𝗲𝗿𝗶𝗻𝗴, 𝗞𝗡𝗡, 𝗚𝗕𝗠, 𝗣𝗖𝗔, 𝗦𝗩𝗗.

• Worked as a Data Science Analyst on pediatric cancer patient data from SIU School of Medicine (SIUSOM). • Applied statistical & predictive modeling techniques to identify significant patterns in the diagnosis and treatment of the SIUSOM’s service region of central and southern Illinois.• Performed the initial statistical exploration using R to gain insights into the data• Analyzed data to find specific correlations between demographic information and disease unique to the pediatric patients that are served by SIUSOM.• Concentrated on linkages between socioeconomic status, racial/ethnic demographics, zip code, and other demographic factors and specific pediatric diseases.• Analyzed these linkages in a manner that allows clinician researchers to reexamine our approaches and treatment plans for the patients that we serve.𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 𝗦𝘁𝗮𝗰𝗸: 𝗥, 𝗦𝗤𝗟, 𝗠𝗦 𝗘𝗫𝗖𝗘𝗟, 𝗧𝗮𝗯𝗹𝗲𝗮𝘂, 𝗘𝘅𝗽𝗹𝗼𝗿𝗮𝘁𝗼𝗿𝘆 𝗗𝗮𝘁𝗮 𝗔𝗻𝗮𝗹𝘆𝘀𝗶𝘀, 𝗖𝗵𝗶-𝗦𝗾𝘂𝗮𝗿𝗲𝗱 𝗧𝗲𝘀𝘁, 𝗔𝘀𝘀𝗼𝗰𝗶𝗮𝘁𝗶𝗼𝗻 𝗥𝘂𝗹𝗲𝘀, 𝗟𝗼𝗴𝗶𝘀𝘁𝗶𝗰 𝗥𝗲𝗴𝗿𝗲𝘀𝘀𝗶𝗼𝗻, 𝗟𝗗𝗔, 𝗤𝗗𝗔, 𝗞𝗡𝗡, 𝘁𝗿𝗲𝗲 𝗰𝗹𝗮𝘀𝘀𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻, 𝗮𝗻𝗱 𝗿𝗮𝗻𝗱𝗼𝗺 𝗳𝗼𝗿𝗲𝘀𝘁 𝗺𝗲𝘁𝗵𝗼𝗱.

• Worked on a Supply Chain Management Project to ensure a proper level of inventory in a warehouse by forecast recommendations.• Collected and reviewed data from different research studies.• Wrote R code to perform statistical analysis and meta-analysis on data to get key insights.• Used R packages like Carat, ggplot2, sqldf, metafor to create statistical models and perform visual data analysis.• Involved in defining the source to target data mappings, business rules, and data definitions.• Performed data analysis and data profiling using complex SQL on various sources systems.• Created dashboards in Excel to understand visualizations for non-technical users. Programming languages\Techniques\tools: R, Excel, Pivot table, VLOOKUP, SQL, t-test, Hypothesis testing, descriptive Statistics.

• Worked on a project name A Novel Computer Aided Diagnosis(CAD) System for Breast Cancer Based on Convolution Neural Network(CNN).• It is presented using an innovative tumor classification method based on convolution neural network on breast mammogram images to improve the diagnosis accuracy.• A novel training and testing scheme is also proposed to train CNN in an image dataset with limited case number. • A novel reinforcement sample learning scheme is proposed to train the CNN with less iteration epochs and less training time.Programming languages\Techniques\tools: MATLAB, R, Neural Network, Conjugate Gradient Method.

• Proposed a new idea where money can be transferred easily in absence or low range of cellular network. • Analyzed survey data on mobile payment gateways in India to ensure the product launching strategy. • Contributed in ensuring all implementations consistent with Software Development Lifecycle (SDLC) standards. • Involved in creating process flow diagrams for overall business specification. • Involved in understanding of data extraction, which involves Online Transfer Processing (OLTP) system and centralized database server. Programming languages\Techniques\tools: R, ETL Processing, SQL, MS ACCESS, MS EXCEL (V-Lookup, Macros, Pivot Tables).

• Worked on a project named Document Processing System, which holds data for customers involved in a claim.