• Designed and applied the CVS Health Medication Adherence AI platform using AWS Bedrock LLMs, developing agentic AI workflows that

autonomously reasoned over patient context, invoked clinical data tools, and generated personalized medication recommendations, collaborating

with clinical and data teams and improving early recommendation accuracy by 8%.

• Built and optimized the patient data ingestion pipeline using AWS Lambda and S3, performing extensive feature engineering on patient records,

cleaning inconsistent values, and enhancing data quality, improving adherence prediction reliability by 12% across datasets.

• Developed, trained, and fine-tuned adherence prediction models using SageMaker, performing hyperparameter tuning for optimal performance,

leveraging Python and Pandas, and containerizing with Docker and Kubernetes for secure, HIPAA-compliant production deployment.

• Implemented validation and testing strategies, including cross-validation and A/B testing, measuring model performance on unseen patient data,

ensuring predictions maintained >95% consistency with internal clinical benchmarks and reducing erroneous recommendations.

• Established monitoring and observability using CloudWatch, Airflow pipelines, and logging frameworks, detecting runtime errors and system

anomalies, reducing downtime by 10%, and maintaining reliable, production-ready generative AI workflows for patient care teams.

• Collaborated with data, cloud, clinical, and DevOps teams to standardize deployment pipelines using MLflow, DVC, and FastAPI, enabling agentic AI

systems capable of multi-step decision-making, pipeline orchestration, and self-triggered model inference, improving operational efficiency by 17%

while ensuring scalable, production-ready deployments.

• Worked on the Zoho Insights Engine, developing LightGBM and CatBoost models to forecast customer behavior, sales trends, and subscription

renewals, team up with product and business teams, improving forecast accuracy by 30% and optimizing revenue operations.

• Designed a automated enterprise data pipelines using Python, SQL, PostgreSQL, BigQuery, Docker, Airflow, MLflow, and DVC, processing large-scale

CRM and SaaS application datasets, enabling continuous model updates, secure deployment, and scalable integration across Zoho applications.

• Developed GNN models for the Zoho Workflow Optimizer Project, optimizing task assignment, workflow automation, and cross-team collaboration

using PyTorch, TensorFlow, AWS SageMaker, FastAPI, and Kubernetes, enhancing productivity and user engagement across enterprise customers.

• Developed Investor Insights Platform by collaborating with portfolio managers, data scientists, and stakeholders during requirement-gathering

sessions and analyzed transactional, account, and market feedback data using BERT and transformers, improving predictive accuracy of client

investment behavior by 10% and enabling actionable portfolio recommendations.

• Built fraud detection models for financial transactions using LightGBM, integrating trading, account, and client behavior data, and coordinated with

compliance and finance teams to reduce fraudulent activities by 7%, strengthening security and operational efficiency across financial operations.

• Designed predictive maintenance models for financial systems and trading platforms by coordinating with IT and operations teams and applying time

series forecasting and anomaly detection, reducing system downtime by 12% and ensuring reliable and continuous investor services.

• Engineered personalized financial product recommendations by applying joint filtering and matrix factorization on client transaction history and

engagement data and working with product and advisory teams, improving adoption, increasing customer retention and enhancing satisfaction.

• Automated end-to-end ML pipelines for financial applications by directing with IT and operations teams, streamlining data ingestion, model training,

deployment, version control, and monitoring, delivering production-ready systems supporting trading, investment advisory, and risk management.

• Implemented monitoring and data integrity pipelines by collaborating with QA and data engineering teams, designing dashboards and automated

alerts to reduce model drift by 5%, ensure financial data accuracy, and maintain full compliance with regulatory standards.