Pranay RS

Fine-tuned biomedical LLMs using PEFT and LoRA on clinical and genomic text, achieving 99% classification accuracy across 50K+ patient records for oncology and translational research workflows. Built and optimized terabyte-scale multimodal pipelines in Azure Databricks and Spark using Data Lake and Blob Storage to process genomic and imaging data for large-scale clinical trial analytics. Optimized retrieval and evaluation pipelines using LangChain and FAISS, improving clinical insight consistency by 20% and reducing response time for precision oncology use cases.

Built 10+ interactive Power BI dashboards with features like dynamic query parameters, improving data visibility and enabling 25% faster operational insights for 50+ stakeholders across engineering and operations. Developed SQL queries and ETL scripts to clean, transform, and integrate 5M+ records, improving the accuracy and consistency of reporting pipelines while partnering with cross-functional teams to align outputs to key KPIs.