Jan 2025 – Present · 1 yr 2 mos · Hyderabad, Telangana, India · Hybrid

• Responsibilities:
• Developed patient and market insights by analyzing treatment preferences, drug-switching patterns,
• and new patient starts (NPS) for alopecia patients, comparing client and competitor drugs.
• Developed HCP segmentation using Rx data and business rules, implementing an ML model on the
• LAAD dataset to predict segment likelihood and identify key drivers differentiating segments. Created
• Sankey visualizations with Plotly to analyze patient journeys, leveraging data-driven insights for strategic
• decision-making.
• Evaluated campaign effectiveness by analyzing pre- and post-campaign impact on disease-modifying
• therapy adoption across zip codes, identifying high-performing and under-performing regions for
• strategic improvements.
• Provided actionable insights for decision-making by uncovering prescribing patterns, optimizing
• patient engagement strategies, and identifying opportunities for market growth

Apr 2022 – Dec 2024 · 2 yrs 8 mos · Hyderabad, Telangana, India · Hybrid

• Responsibilities:
• Led and enhanced the existing Recommendation model and introduced the ALS model for content recommendation. Address the cold start problem using the user and item embeddings, improving overall system performance.
• Designed and implemented scalable and efficient machine learning models on Amazon SageMaker, leveraging advanced algorithms and best practices to deliver high-performance solutions tailored to client needs.
• Established rigorous unit testing frameworks and methodologies to validate the functionality and reliability of machine learning code deployed on Amazon SageMaker, ensuring robustness and stability of the production environment.
• Analyzed the OncEmr data to create patient cohort and prepared datasets for TNM Classifier and Supported MLOps in integrating data science code into production on AWS.
• Designed an unsupervised learning methodology to mine trends related to specific need states in the skincare category by building Knowledge Graphs on entities extracted from social media posts using a span categorizer. Developed clustering technique and used statistical approaches for outlier detection.
• Ingested data from YouTube using APIs for time series and sentiment analysis on Cerave Skincare comments. Applied NLP techniques like NER and Topic Modeling to extract key insights.
• Communicated project progress, insights, and findings to clients regularly, ensuring transparency and alignment with goals. Led internal meetings to discuss updates, address concerns, and propose solutions. Ensured smooth collaboration between teams and stakeholders for effective decision-making.
• Key Achievements:
• Successfully tuned the Light FM model, achieving a 22% improvement in recommendation accuracy.
• Implemented an alternative recommendation solution using the ALS model, effectively addressing the cold start issue through custom functionality with embeddings and client acceptance of the recommendations.