My name is Haniye Kashgarani a Senior AI Scientist in RCAC. My background is in CS, and I specialize in high-performance computing, artificial intelligence, algorithm selection and combinatorial optimization.

I am a Senior AI Scientist on the Applications team at RCAC, where I focus on strengthening RCAC's AI ecosystem by improving documentation, datasets, and workflows to help researchers use advanced computing resources more effectively. I provide AI consultations, deliver AI-focused trainings, and support NAIRR-related activities. I also collaborate on research projects that require AI expertise.

You would come to me for help because I enjoy problem-solving and approach challenges with curiosity. With my experience in different computer science sub fields (i.e. software development, AI, ML, HPC, GPU computing), research, and mentoring, I try to provide guidance for specific situation.

I joined RCAC in April 2024. One thing I wish I’d known when I started working in HPC is just how much of the ecosystem depends on understanding the layers around the compute itself—Kubernetes, cloud workflows, storage systems, and all the operational pieces that support AI pipelines.

I’m currently working on becoming more familiar with modern development processes so I can navigate the infrastructure side more confidently. Strengthening these skills helps me focus more on applicable research and supporting teams with end-to-end AI solutions.

My go-to productivity trick is a mix of digital note-taking and reminders along with journaling to stay organized and mentally clear.

‘Success’ is a relative word, but I think whatever success I’ve had comes from being curious, flexible, and eager to try new things. I enjoy getting involved in different projects and activities because it helps me learn and grow. Sometimes I feel a bit lost, but my excitement to keep learning and improving always pushes me forward.

After my undergraduate studies, my first job was as a .NET developer for a company that worked on a hotel booking website (2017).

My favorite job is my current role. It allows me to blend hands-on technical work, research, and meaningful collaboration—supporting researchers, mentoring students, and helping teams solve complex challenges. I also appreciate that the role continually pushes me to grow, letting me contribute to projects that genuinely improve access to advanced AI and computing resources.

I’m excited about mentoring students in the REU program and helping them explore the world of HPC and AI. It’s incredibly rewarding to guide them and watch their progress. I also enjoy working with researchers from many different fields—seeing their diverse applications and scientific goals really broadens my perspective and keeps the work inspiring.

I was pleasantly surprised by the strong sense of collaboration and support among my colleagues at RCAC. Everyone is so knowledgeable and willing to help, which creates a fantastic learning environment.

The title ‘AI Scientist’ can sound a bit confusing—people often assume it’s narrowly focused, but the role is actually very diverse. It includes consulting, providing application-level infrastructure support, and contributing to a wide range of initiatives and research projects.

I started in April 2024!

I earned my B.S. in Software Engineering from Shariaty Technical College of Tehran, Iran, where I specialized in .NET development. However, after starting my PhD, I transitioned away from .NET development and focused on research in combinatorial optimization and high-performance computing (HPC). During my PhD at the University of Wyoming under the supervision of Dr. Lars Kotthoff, my research centered on optimizing combinatorial problem-solving through algorithm selection and parallel subportfolio selection using machine learning and problem features. My experiments required solving computationally intensive problems like Boolean satisfiability, which led to extensive use of HPC clusters for running large-scale experiments. I also completed two internships at NCAR:

• Developing CI/CD Pipelines: Built and tested GPU-parallelized code for NCAR developers to streamline their workflows.
• Geoscience Data Analysis Toolkit: Ported a Python-based toolkit from CPU parallelization to GPU to handle the computational demands of geoscience workflows.

Throughout my academic journey, I’ve been deeply engaged with the HPC community, using and contributing to cutting-edge tools and methods. See my Publications.

 

I love playing the electric guitar —“a beginner today, a master someday :-D”— teaching my dog Jane new tricks, and watching movies. 

I decided to become a mentor for the REU program because I enjoy problem-solving and helping others grow. Having benefited from mentorship myself during my PhD and internships, I know how valuable guidance and support can be. Mentoring REU students is an exciting opportunity to share what I’ve learned in HPC and research, and it’s incredibly rewarding to watch students explore their potential and build confidence in their abilities.

Being a mentor for the REU program is incredibly rewarding. It gives me the chance to share my knowledge and experiences, but just as importantly, I learn from the fresh perspectives and curiosity the students bring. Every mentorship experience teaches me something new.

Your REU students will work on building a multi-agent AI system that supports scientists through the full research lifecycle — from reading papers, to preparing datasets, to running experiments, to ensuring the results are reproducible. They’ll develop components that directly address some of the biggest challenges in modern scientific research: information overload, inconsistent documentation, and difficulty reproducing results.

In the next decade, I see HPC and supercomputing evolving to become more accessible and integrated into various disciplines. With the rise of artificial intelligence and machine learning, we’ll likely see HPC systems more optimized for data-intensive applications and real-time processing. Optimization of any kind will also impact energy efficiency and sustainability.

No