Dhairya Bhatt

Building automated pipelines to collect, process, and analyze research papers from top-tier journals across medicine, information systems, and engineering. Developing text extraction and NLP-based classification systems to detect AI acknowledgment patterns in published work, supporting the research team's control-treatment study design. Working under Prof. Anand Nair and Prof. Guohou Shan.

Mentoring new graduate students through their transition to Northeastern by providing one-on-one guidance, connecting them with campus resources and support services, and helping them build community and belonging both on campus and around Boston.

CS 3650: Computer Systems - Supported 300+ students as a TA for Systems Programming by conducting weekly office hours and hands-on debugging sessions for large-scale C/C++ projects involving assembly, cache optimization, virtual memory management, and CUDA kernels, ensuring performance efficiency. - Diagnosed and resolved memory leaks, race conditions, and performance bottlenecks while mentoring students on pthreads, mutex/semaphore synchronization, system calls, and GPU parallelism, improving execution efficiency by 20% and scalability across intensive workloads.

As a SAP UI5/Fiori Developer, I led the end-to-end development and customization of enterprise-grade Fiori applications, translating complex functional requirements into clean, responsive, and user-centric interfaces. I optimized frontend performance through adaptive UI logic, modular component design, and efficient OData integrations, reducing repetitive code by 10% and improving rendering speed for 80+ internal users. My work streamlined high-volume business workflows, enhanced usability, and contributed to faster task completion across teams.

• Engineered PatchIQ, a Content Based Image Retrieval (CBIR) system for satellite payload validation, processing 590K BigEarthNet hyperspectral images across 13 spectral bands to automate ISRO's manual testing workflows. • Implemented deep metric learning with PyTorch using contrastive loss and triplet networks for robust visual embeddings, improving retrieval accuracy from 66% to 89% on large scale remote sensing datasets. • Architected ETL pipelines integrating locality sensitive hashing (LSH) and hard negative mining strategies, achieving sub-second semantic retrieval across the entire satellite imagery dataset. • Built an interactive Next.js dashboard for real time query visualization and embedding analytics, enabling scientists to explore similarity rankings and retrieval patterns. • Presented findings to 50+ senior ISRO scientists, receiving approval for deployment in the operational payload validation pipeline.

Built scalable full-stack web applications using React.js, Node.js/Express, and PostgreSQL architecture. Key contributions: - Designed decoupled frontend-backend architecture with lazy loading and responsive UIs, cutting page load times by 40% - Implemented Test-Driven Development using Jest and React Testing Library for reliable code quality - Automated CI/CD pipelines with GitHub Actions and Docker, achieving 90% test coverage and daily releases - Enhanced security with JWT authentication, OAuth 2.0, bcrypt encryption, and parameterized SQL queries

Relevant Coursework: - Programming Design Paradigm (CS 5010) - Essentials of Data Science (DS 5110) - Algorithms (CS 5800) - Foundations of Artificial Intelligence (CS 5100)

• Relevant Coursework : Machine Learning, Deep Learning, Artificial Intelligence, Full Stack Development, Object-Oriented Programming, Operating Systems, Advanced Java Technologies (JEE, Spring, SpringBoot, Hibernate), Distributed Computing (Docker, Kubernetes), Data and File Structures, Algorithms, E-Commerce and E-Security, Data Mining and Data Warehousing, Database Management System, Computer Networks, Theory of Computation, Microprocessors, Communication Systems, Compiler Design, Algebra, Calculus, Graph Theory, Probability