Kavisha Ghodasara

Under the guidance of Professor Andrew McCallum, I’m working on a collaborative research project with Microsoft that explores diffusion-based approaches to modeling memory, focusing on how information naturally fades and can be recalled within intelligent systems.

Serving as a grader and teaching assistant for COMPSCI 520: Theory and Practice of Software Engineering at the University of Massachusetts Amherst, supporting instruction and assessment throughout the semester. Leveraged deep understanding of software development principles to evaluate student work, provide detailed feedback, and assist with course facilitation and student queries.

- Designed a pharma instrument scheduling optimizer using MILP (PuLP/CBC), streamlining 100+ daily tasks across 15-day horizons while maintaining full regulatory compliance. - Built a multi-worker parallel pipeline to process 10K+ records, cutting computation time by 93% and boosting solver performance up to 5×. - Developed a constraint validation engine with real-time conflict detection and expertise-based prioritization, ensuring accuracy and compliance across all scheduling operations. - Collaborated with cross-functional teams to interpret manufacturing data trends, identify inefficiencies, and propose data-driven improvements in production workflows.

- Led a team of four developers to build a modular, UI-driven cloud management solution integrating Microsoft APIs for Intune, Windows 365, and license management—boosting platform automation and security by 75%. - Re-architected SQL schemas and implemented MySQL Query Cache, improving query efficiency and reducing response times by 30–40%. - Delivered 30+ Angular screens to streamline tenant onboarding/offboarding, cutting setup time by 30% and improving client satisfaction metrics. - Engineered client-side pagination and optimized data retrieval pipelines, reducing API payload sizes and improving load times by 70%. - Built a real-time observability dashboard for 5,000+ managed devices, enhancing data transparency and system diagnostics for enterprise customers. - Drove Agile practices through sprint planning and retrospectives, ensuring consistent delivery across multi-module releases.

- Enhanced a campsite booking platform by integrating multilingual translation, dynamic QR-code check-ins, and discount-driven billing logic, leading to 20% higher conversions. - Designed a smart billing engine to dynamically apply location- and duration-based discounts, improving payment transparency and user retention. - Implemented third-party login and authentication for a seamless customer experience, reducing login time by 90%. - Gained early exposure to Docker containerization, backend APIs (Django REST), and collaborative version control practices.

Courses taken: PhD level courses: COMPSCI 602: Research Methods in Empirical Computer Science COMPSCI 666: Theory and Practice of Cryptography COMPSCI 682: Neural Networks: A Modern Introduction COMPSCI 683: Artificial Intelligence COMPSCI 698DS: Practicum - Artificial Intelligence DACSS 601: Data Science Fundamentals Grad level: COMPSCI 520: Theory and Practice of Software Engineering COMPSCI 561: System Defence and Test STAT 501: Math Applied Stats

- Graduated with an honors degree in Data Science - Relevant Coursework: Data Structures and Algorithms, Object Oriented Programming, Information theory coding and Computer networks, System Programming and Operating Systems, Database Management, Artificial Intelligence(AI), Machine Learning(ML)