Shishmitha Adusumilli Email & Phone Number

San Francisco Bay Area

>> 𝐀𝐈 𝐬𝐡𝐚𝐫𝐞𝐝 𝐬𝐞𝐫𝐯𝐢𝐜𝐞𝐬 𝐟𝐨𝐫 𝐀𝐈 𝐞𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠• Contributed to the development of generative AI applications offering both RPC and REST APIs leveraging Azure OpenAI and Azure Document Intelligence (ADI) services, improving operational efficiency by ~42%.• Designed and developed an AI Software Development Kit (SDK) using the LangChain framework for the ADI service, streamlining the creation of AI-powered tools like chatbots and data analysis applications.• Automated secrets management for the applications by integrating HashiCorp Vault with the Azure Secrets Engine, optimizing credential handling and enhancing security.>> 𝐑𝐞𝐥𝐨𝐚𝐝 𝐭𝐫𝐚𝐝𝐞𝐬 𝐀𝐏𝐈 𝐟𝐨𝐫 𝐄𝐓𝐅 𝐬𝐞𝐫𝐯𝐢𝐜𝐢𝐧𝐠 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦 (𝐄𝐒𝐏)• Developed a data-efficient, on-demand reload trades API to rectify erroneous trades within the ETF servicing platform (ESP), ensuring the timely generation of accurate ETF basket data for daily operations and handling $3 trillion in assets under management (AUM) for the iShares ETF business, including the pioneering Bitcoin ETF.• Validated the reload trades interface’s reliability and scalability through developing a meticulous test plan and performing rigorous functional testing.>> 𝐃𝐞𝐫𝐢𝐯𝐚𝐭𝐢𝐯𝐞 𝐜𝐨𝐥𝐥𝐚𝐭𝐞𝐫𝐚𝐥 (𝐃𝐄𝐂𝐎) 𝐰𝐞𝐛 𝐚𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐟𝐨𝐫 𝐜𝐨𝐥𝐥𝐚𝐭𝐞𝐫𝐚𝐥 𝐦𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭• Led the development and launch of DECO, an Angular-based web application for managing and updating collateral statements.• Collaborated with the product team to implement key user-driven enhancements, resulting in a 55% increase in engagement.

Madison, Wisconsin, United States

• Designed a data scraping approach to extract structured content from scientific articles for NLP applications.• Developed APIs to efficiently retrieve and parse scientific papers based on keyword searches, enhancing data acquisition capabilities from prominent web databases.

Department Of Computer Science, Madison, Wi

CS 354 Machine Organization and ProgrammingCS 540 Artificial Intelligence• Assisted students with projects and homework assignments during office hours.• Mentored students by directing them towards helpful resources and conducting debugging sessions.

• Evaluated state-of-the-art object detection models and trained Cascade R-CNN model using advanced deep learning techniques, significantly enhancing defect detection in metals and cancer diagnosis from medical images.• Achieved a ~75% reduction in detection costs and a ~38% improvement in accuracy with Cascade R-CNN, directly contributing to improved industrial efficiency and early-stage medical diagnoses.

• Assisted students 1-on-1 in comprehending Java programs.• Provided references to online resources like Codecademy, GeeksforGeeks, etc. to improve students' coding skills while encouraging them to be an independent programmer.

Madison, Wisconsin, United States

• Developed a QR code application in Node.js that ensured secure handling and storage of sensitive data, reducing security software costs by 25%.• Implemented encryption and authentication protocols, and deployed across Microsoft Azure and AWS for cross-cloud flexibility.

• Supported 60-80 students during the weekly discussions by providing handouts of practice problems and conducting quizzes.• Helped students build their skills and knowledge in the course material by assisting them with online assignments and homework.