Shanmukha Sai Dheeraz Chavali

• Research and Teaching Assistant under Prof. Dr. Lukas Burkart, spearheading interdisciplinary AI-humanities research to integrate LLMs and generative models into historical scholarship, transforming centuries-old texts into structured, machine readable formats. • Leveraging advanced Large Language Models (LLMs) to digitize and structure scholarly texts of ancient historians, improving online accessibility and usability.

• Designed and deployed a fully customized Amazon Connect instance integrated with Salesforce via Lambda, reducing third-party licensing dependency and saving the client $40,000/quarter in operational costs. • Designed a predictive analytics pipeline over voice call logs and voicemails using S3, Glue, Redshift, Sage Maker, enabling real time concern clustering and saving $60,000/quarter through smarter call routing and resource allocation.

• Streamlined a hash-based skew mitigation algorithm in Spark, improving large join performance by 30% in fraud detection workloads. • Refactored inefficient ETL jobs and restructured S3 storage classes with intelligent tiering, cutting monthly storage and compute expenses by $25,000 while improving data retrieval SLAs. • Amplified graph-based customer segmentation models using community detection, hierarchical clustering, increasing campaign ROI by 45%.