Biography

Ashna Nawar Ahmed is a 2nd year Ph.D. student in Computer Science at Texas State University, where she works with Dr. Tanzima Islam in the Per4ML Research Group. Her research focuses on applying machine learning and surrogate modeling techniques to optimize performance and energy trade-offs in high-performance computing (HPC) systems. She explores intelligent, data-driven approaches for multi-objective optimization, scheduling, and sustainable computing. Ashna’s recent paper, “Attention-Informed Surrogates for Navigating Power-Performance Trade-offs in HPC,” was accepted to the Machine Learning for Systems (MLForSys) Workshop at NeurIPS 2025, and her poster “Intelligent Surrogates Pay Attention to Data” will be presented at SC25. She also completed a Graduate Research Internship at Oak Ridge National Laboratory (ORNL) under Terry Jones, contributing to scalable optimization for HPC systems. Before pursuing her doctorate, Ashna served as a Lecturer at Ahsanullah University of Science and Technology in Dhaka, Bangladesh, where she graduated first in her class. She has been recognized with multiple academic honors, including an invitation to join the Honor Society of Phi Kappa Phi. Her technical expertise spans Python, PyTorch, BoTorch, DeepSpeed, MPI, CUDA, and large-scale multi-objective optimization frameworks.

Academic Status

PhD Student - 2nd

Research Area/Department

Computer Science; Machine Learning/AI

Major/Specialty

Computer Science, focusing on High Performance Computing (Scalable systems)

Degrees Earned or in Progress

PhD in Computer Science (2nd Year)

Academic Preparation

I have completed several graduate-level computer science courses that have equipped me with both theoretical and practical foundations for a summer internship. These include CS 7331 – High-Performance Computing, where I gained hands-on experience with parallel programming, distributed computing frameworks, and performance optimization on large-scale systems; and CS 7315 – Network Science, which deepened my understanding of complex systems, graph analytics, and data-driven modeling. Additionally, CS 7334 – Scalable Systems strengthened my skills in designing and managing systems that handle large workloads efficiently, while CS 7389F – Secure Cyber-Physical Systems introduced advanced topics in system security and resilience. I also completed CS 7300 – Introduction to Computer Science Research, which enhanced my ability to conduct independent research. Together, these courses have provided me with the computational, analytical, and research skills necessary to contribute effectively to an internship in areas such as machine learning, HPC systems, or large-scale data analytics.

Research/Publications

My research focuses on machine learning and surrogate-assisted optimization for high-performance computing (HPC) systems. I have conducted this work as a Doctoral Research Assistant in the Per4ML Research Group at Texas State University, under the supervision of Dr. Tanzima Islam, and during a Graduate Research Internship at Oak Ridge National Laboratory (ORNL) with Terry Jones. My research has been published and presented in leading venues in the HPC and machine learning communities: “Attention-Informed Surrogates for Navigating Power-Performance Trade-offs in HPC”, Machine Learning for Systems (MLForSys) Workshop, NeurIPS 2025. [Accepted] “Intelligent Surrogates Pay Attention to Data, Improving Multi-Objective HPC Optimization”, Supercomputing Conference (SC25) poster. [Accepted] “Decentralizing and Optimizing Nation-Wide Employee Allocation While Maximizing Employee Satisfaction”, Lecture Notes in Networks and Systems (Springer, 2022) — DOI: 10.1007/978-3-031-22039-5_17

Research/Academic Interests

My research focuses on applying machine learning and optimization techniques to high-performance computing (HPC) systems. I am particularly interested in developing intelligent surrogate models that can efficiently explore trade-offs between performance, energy consumption, and scalability in large-scale computing environments. By combining multi-objective Bayesian optimization with attention-based neural networks, my work aims to improve decision-making in HPC job scheduling and resource allocation for sustainable computing. As part of the Per4ML Research Group at Texas State University, under Dr. Tanzima Islam, I am currently building embedding-informed surrogate frameworks that integrate deep learning with traditional optimization to enhance predictive accuracy and generalization across HPC datasets. My broader academic interests include AI for systems, energy-aware computing, and evolutionary algorithms, with the long-term goal of enabling more efficient and environmentally responsible large-scale computing systems.

Computational and Data Science Areas

Applied Computer Science; Artificial Intelligence and Intelligent Systems; Computer Science; Informatics, Analytics and Information Science; Performance Evaluation and Benchmarking

Motivation

As a Ph.D. student in Computer Science at Texas State University and a member of the Per4ML Research Group, my research lies at the intersection of machine learning and high-performance computing (HPC), specifically, in developing attention-informed surrogate models that optimize the balance between performance, energy efficiency, and scalability in large-scale computing environments. This focus on data-driven, sustainable computing has inspired me to seek opportunities that not only deepen my technical expertise but also allow me to engage with diverse scientific communities committed to inclusive innovation. The Sustainable Research Pathways (SRP) program perfectly aligns with these goals. The mission of the Sustainable Horizons Institute, which is: to build inclusive, sustainable scientific ecosystems where everyone can thrive, resonates deeply with my own academic philosophy. Throughout my journey from Bangladesh to the United States, I have witnessed how access, mentorship, and community support can transform a student’s trajectory. I am passionate about contributing to such an environment where researchers from all backgrounds can find a sense of belonging while advancing impactful science. By participating in SRP, I hope to not only strengthen my technical and research capabilities but also to learn how to translate inclusivity into everyday research practice, through mentorship, collaboration, and shared scientific purpose. My research on multi-objective optimization and machine learning for HPC scheduling directly complements the objectives of the NSF National Artificial Intelligence Research Resource (NAIRR) initiative. In my recent work, I proposed attention-informed surrogate models that integrate deep learning and Bayesian optimization to efficiently navigate runtime–power trade-offs in HPC systems. This work, accepted to the Machine Learning for Systems (MLForSys) Workshop at NeurIPS 2025 and as a poster at SC25, demonstrates my commitment to sustainable and scalable AI-driven computing. Through the SRP program, I hope to extend this research by leveraging NAIRR resources and collaborating with domain experts to explore cross-disciplinary applications of intelligent optimization. Beyond the technical benefits, I am drawn to SRP’s emphasis on community-building and mentorship. I have experienced firsthand how mentorship from my advisor, Dr. Tanzima Islam, and collaborators at Oak Ridge National Laboratory (ORNL) has shaped my confidence as a researcher. I hope to pay that forward by learning how to mentor and support others, particularly, underrepresented students, through structured, sustainable programs like SRP. I believe that creating equitable pathways into research not only strengthens individual careers but also leads to more robust, creative, and socially responsible science. Through this program, I aim to gain practical experience working on real-world NAIRR-aligned projects, expand my professional network, and develop the interdisciplinary communication skills required to thrive in collaborative scientific environments. More importantly, I want to grow as a researcher who not only builds intelligent systems but also helps build the communities that sustain them. In essence, the SRP program represents an ideal opportunity to combine my technical expertise in AI for sustainable computing with my commitment to inclusive research ecosystems, a combination I believe is essential for the future of science and technology.