🧭 Perspective & Professional Insight 🧠 "In my opinion, there are two distinct paths in High-performance computing (HPC): one leaning toward system administration, and the other toward scientific software engineering."
Important: This matrix is intended for professionals who will be engaged in tasks related to on-premise HPC initiatives for the University's facility. For those involving Microsoft Azure Cloud, I recommend taking the Azure AZ-104 certification: The Azure Administrator Associate.
The system administration path emphasizes foundational skills in Linux, networking, and hardware management (CPU, GPU, FPGA). The scientific software path leans into coding, algorithms, and specialized HPC frameworks and tools. Having worked as a Scientific Software Engineer for several years, I designed this matrix using Bloom’s Taxonomy to reflect the skill progression I’ve observed and experienced firsthand. It integrates over 25 years of deep technical involvement in Linux, networking, and hands-on coding, as well as the advanced tools and workflows I utilized across roles. This framework is intended to help others:
- 🧩 Identify which skills they currently have and at what level
- 🔍 Choose where to invest learning time based on interest or career goals
- 📈 Advance from foundational (Level 1) to expert-level mastery (Level 6)
For example:
- Skills like Linux and Python often provide a solid foundation for HPC System Engineering.
- Specialized technologies and frameworks mark progression into Scientific Software Engineering. The more skills you master and elevate toward higher cognitive levels — especially those aligned toward the right columns of the matrix — the deeper your expertise and seniority in HPC becomes.
This repository provides a structured overview of technical competencies and cognitive development across High Performance Computing (HPC) domains. Based on Bloom's Taxonomy, it helps assess skill levels for HPC professionals and guide learning paths.
| Level | Process | Example Verbs | Description |
|---|---|---|---|
| 1 | Knowledge | define, list, recall | Retrieve basic facts or knowledge |
| 2 | Comprehension | explain, summarize, classify | Explain ideas or concepts |
| 3 | Application | use, implement, solve | Use knowledge in new situations |
| 4 | Analysis | compare, differentiate, test | Break down to explore patterns and relationships |
| 5 | Synthesis | critique, judge, justify | Make judgments based on criteria |
| 6 | Evaluation | design, construct, invent | Create new ideas, products, or perspectives |
| Level | Description |
|---|---|
| I - Basic | Foundational knowledge and tool awareness |
| II - Intermediate | Hands-on experience and partial autonomy |
| III - Advanced | Confident execution and troubleshooting |
| IV - Specialized | Deep expertise and innovation-driven approach |
- Fortran, CUDA, Perl, CI/CD, C, C++, Java, Python, Shell, MATLAB
- Python & Conda Environments
- LAMMPS, CESM, Alphafold3, GROMACS, OpenFOAM
- Puppet, Chef, Salt, Ansible
- Spack, EasyBuild, dnf, apt
- Slurm scripting & admin commands, job scheduling
- Intel SYCL, OpenMP, Intel VTune, NVIDIA HPC SDK
- Azure, AWS, GCP
- DockerHub, Podman, Singularity, Kubernetes
- Azure-specific tools: AMLFS, Blob Storage integration
- git / GitHub / Bitbucket / SVN
- make, cmake, Snakemake, NextFlow
- PyTorch, TensorFlow, JAX
- Feature Engineering, Large Language Models (LLMs)
- Agentic Systems, Retrieval-Augmented Generation (RAG)
- Linux System Administration (LPIC-1/2)
- iptables, port & service management
- Performance profiling and tuning
- Jupyter, VSCode, RStudio, OpenOnDemand
- Confluence Web Server, XDMoD, CryoSparc, Coldfront
- Quantum Computing, Supabase, SBGrid, warewulf, MAAS
| Role | Target Skill Level |
|---|---|
| HPC Scientific Software Engineer | II – III |
| HPC Sr. Scientific Software Engineer | III – IV |
For a comprehensive visual representation of the HPC skills matrix based on Bloom's Taxonomy, please refer to:
This PDF provides a detailed matrix mapping HPC technologies and skills across cognitive levels and skill progression paths, based on my profile when I worked at The Advanced Research Computing at Hopkins (ARCH) HPC facilities, 2021-2025.