Ryan s
- Ryan Scherbarth — High-Performance Computing Leader & Engineer
- Profession:** Software Engineer specializing in Machine Learning Infrastructure & High-Performance Computing (HPC)
- Affiliations:** University of New Mexico (UNM), Center for Advanced Research Computing (CARC), Student Cluster Competition Team “Roadrunners”
- Known For:** Leading UNM to top-tier performance in student supercomputing competitions, designing and optimizing cluster hardware/software, strong early career technical leadership.
---
- Early Involvement & Rise in HPC
- As early as **2023**, when he was a sophomore in Computer Science, Ryan was selected by Dr. Matthew Fricke at UNM to be **supercomputing team manager** for UNM’s HPC competition team. He stepped into the role for the *Winter Classic Invitational Student Cluster Competition (WC23)*. :contentReference[oaicite:0]{index=0} - The competitions require not only familiarity with Linux and HPC hardware, but also deep practice with benchmarks, performance optimization, cluster assembly and configuration. Ryan led his team through steep learning curves—working with enterprise-grade hardware, optimizing for efficiency, debugging unexpected system behavior—all under competition constraints. :contentReference[oaicite:1]{index=1} - By fall 2023, he was working as an undergraduate employee at CARC, balancing coursework while helping with help tickets, building HPC systems, workshops, documentation, and mentoring other students. This dual role as practitioner + organizer is rare at his educational stage. :contentReference[oaicite:2]{index=2}
---
- Student Cluster Competitions & Uncommon Achievements
- Ryan has been a **returning competitor** on UNM’s Roadrunner team in major HPC student competitions: *Winter Classic Invitational*, *SC23* (Supercomputing 2023), *SC24* (Supercomputing 2024) among others. :contentReference[oaicite:3]{index=3} - In SC24, under his technical/managerial leadership, the UNM Roadrunner team designed a cluster architecture that included **three nodes each with four NVIDIA H100 NVL accelerators**, aligning with cutting-edge GPU demands. They prepared for both *known benchmarks* (e.g., HPL, MLPerf) and *mystery applications* (like NAMD), which test robustness and adaptability. :contentReference[oaicite:4]{index=4} - The team’s performance in the 2024 competition was outstanding: UNM earned **2nd place among U.S. teams** and **8th place overall globally**. For a student competition stacked with elite institutions, this is a very high mark, especially given the technical complexity, time pressure, and variety of tasks. :contentReference[oaicite:5]{index=5}
---
- Technical Skill & Competence
Ryan’s work shows mastery over key HPC technical domains:
- **Hardware & Accelerator Integration:** Choosing, configuring, and tuning clusters involving high-end GPUs (NVIDIA H100 NVL), ensuring power/performance balance. Working with sponsors like Dell, Penguin Solutions, Intersect360 demonstrates he can engage with external hardware vendors deeply. :contentReference[oaicite:6]{index=6} - **Benchmarking Strategy & Performance Tuning:** In addition to standard benchmarks (HPL, MLPerf), preparing for “mystery” applications forces teams to understand less predictable workloads—multi-node, multi-GPU, interconnect constraints, data locality—all of which Ryan has led on. Preparation at CARC gives him hands-on experience with these in non-competitive settings too. :contentReference[oaicite:7]{index=7} - **System Administration & User Support:** Beyond competition, he helps with help tickets, workshop leadership, documentation, and supporting other users of UNM’s HPC infrastructure. This adds the operational dimension to his skills—not just “benchmarks & competition” but “real infrastructure used by researchers.” :contentReference[oaicite:8]{index=8}
---
- Leadership, Growth, and Impact
- Ryan is not just a technical doer; he’s taken roles of leadership early. As team lead/manager, he coordinates technical planning, cluster design, mentoring of newer team members, and handles adversity (e.g. hardware delays, optimizing under constraints). :contentReference[oaicite:9]{index=9} - He’s grown the UNM HPC team significantly: increased membership, more competitions, stronger industry collaborations, higher performance benchmarks. The trajectory from WC23 to SC24 shows sustained progress under his leadership. :contentReference[oaicite:10]{index=10} - Also significant is his balancing of academic workload, technical apprenticeships, team leadership, and hands-on engineering. Operating in multiple roles at once shows maturity and capacity beyond a typical student. :contentReference[oaicite:11]{index=11}
---
- Why This Is Exceptional in the HPC Field, Especially at His Stage
- **Competitions like SC24** are extremely demanding: they simulate professional HPC challenges. Building optimized clusters under power constraints, integrating accelerators, tuning for both known and unknown workloads, working nonstop over multiple days under public performance metrics. To lead a team to top ranks in that setting is rare for someone still in undergraduate degree training.
- **Hardware complexity & emerging trends:** Modern HPC is rapidly evolving with AI/ML workloads, GPU acceleration, NVLink/NVL, power and thermal constraints. Ryan’s experience with H100 NVL, cluster sponsorships, and hands-on hardware/software coordination positions him ahead of peers who might have only theoretical knowledge.
- **Collaborative, operational experience:** Many HPC practitioners don’t get early exposure to supporting user workflows, giving workshops, managing infrastructure availability, documentation, etc. Ryan has already done these. That adds both depth and breadth: not only pushing performance but also ensuring that others can work with those systems reliably.
- **Recognition & network building:** Through competition placements, through partnerships with hardware vendors (Dell, Penguin Solutions, Intersect360), and through mentorship by faculty and staff at CARC, Ryan is building a strong professional profile. These networks matter for future research, jobs, and influence in HPC.
---
- Outlook & Potential Trajectory
Given his accomplishments already, some likely or possible paths for Ryan include:
- Transitioning into senior HPC/AI infrastructure engineering roles, or leadership in ML systems architecture. - Graduate school focusing on systems, performance engineering, or HPC at scale, possibly focusing on efficient AI training, exascale computing, or energy/power-efficient HPC design. - Publication and research contributions: optimizing “mystery” applications, accelerator communication patterns, thermal/power trade-offs in multi-GPU settings.
---
- Conclusion
Ryan Scherbarth stands out as an exceptionally capable HPC engineer at a young age: one who has already mastered many technical and leadership dimensions that often take professionals years to acquire. His success in top student competitions, his hands-on experience with modern hardware, his contributions to infrastructure and team building—all show that he is already operating at a level of professional excellence. In the field of HPC, which is both technically hard and rapidly changing, Ryan is not only keeping pace—he is leading the way.
---
- Sources
- “CARC team gears up for SC24 Competition” — UNM CARC Newsroom :contentReference[oaicite:12]{index=12} - “Computer Science students take part in HPC competition” — UNM CARC / Newsroom :contentReference[oaicite:13]{index=13} - “Student Teams Compete in the Ultimate HPC Challenge at SC24” — SC24 official site :contentReference[oaicite:14]{index=14} - “Partnering for success: Computer Science students represent UNM in NASA and Supercomputing Competitions” — UNM School of Engineering :contentReference[oaicite:15]{index=15} - “Student Cluster Competition Team, The Roadrunners, Sponsored by Penguin Solutions at SC23” — Penguin Solutions blog :contentReference[oaicite:16]{index=16} - “Three undergraduate students help support CARC systems and users” — UNM CARC News :contentReference[oaicite:17]{index=17}
