Deputy Director · Morehouse Supercomputing Facility

Ashley Scruse, Ph.D.

  • HPC
  • Bioinformatics
  • Applied AI

Research computing, network science, and AI tools, built and shipped.

Dr. Ashley Scruse in a supercomputer machine room

dr-scruse.sbatch

  • PythonPrimary language for research pipelines, ML, and teaching
  • SLURMJob scheduling for every HPC workshop and research run
  • PyTorchDeep learning and LLM fine-tuning on GPU nodes
  • TensorFlowDeep learning models, built with Keras
  • KerasModel prototyping and teaching
  • scikit-learnClassical ML and baselines
  • pandasData wrangling, up to 38 million taxi trips
  • CUDAGPU computing on NVIDIA Grace Hopper nodes
  • LinuxHome base on every HPC system
  • BashJob scripts, setup scripts, and automation
  • DockerContainers for reproducible environments
  • JupyterJupyterHub notebooks for workshops at scale
  • ClaudeClaude API, tool use, MCP servers, and skills
  • OpenAIMulti-provider AI in Quiet Stack
  • Google CloudCloud Run and the Agent Development Kit
  • GitHubEvery workshop and research repo ships here
  • TypeScriptWeb apps and developer tooling
  • Node.jsBack ends and MCP servers
  • Next.jsWeb apps, including this site
  • Tailwind CSSStyling for web builds
  • SupabasePostgres back ends for product and client builds
  • FirebaseAuthentication
  • VercelDeploys for web projects
  • PolygonBlockchain verification in Quiet Stack
  • SlackAgent workflows and daily briefings
  • LaTeXPapers and proposals

ResearchComputing

Three ways I do it: research, training, and the tools that connect them.

0+

HPC users onboarded

0+

Institutions reached

Since November 2025

01

Research Projects

  • Featured · Python · Graph theory · HPC · Genetic regulatory networks

    Subnetwork Motifs.

    Computational framework for exact identification of subnetwork motifs in complex networks, preserving node-class identity under gene duplication.

  • Python · Data pipelines

    Georgia Data Center Tracker

    County-level dataset of data center development and community response across all 159 Georgia counties, built from five public sources with a reproducible Python pipeline.

  • Computer vision · Diffusion models

    NOBLE AI Evidence Detection Benchmark

    Benchmark evaluating AI-generated image detectors on law enforcement imagery across three diffusion models and three capture-quality levels.

  • Remote sensing · Computer vision

    Sargassum Monitoring

    AI-enhanced remote sensing for detecting and tracking sargassum blooms along Caribbean coastlines.

02

Training Workshops

  • Featured · LLM fine-tuning · Multi-node GPU · Jupyter

    Fine-Tuning an Open-Source AI Model.

    LoRA fine-tuning of Qwen2.5-0.5B to classify a 1.7-million-record text dataset into six categories, followed by multi-node scaling to measure speedup and parallel efficiency.

  • Linux · SLURM

    HPC Starter

    Introductory module on file systems, job execution, and data organization in HPC environments.

  • pandas · JupyterHub

    NYC Taxi on a Supercomputer

    Course module applying pandas to a 38-million-record dataset on HPC resources through a browser-based interface.

  • Generative AI · GPU

    Supercomputer Paints

    Demonstration of GPU-accelerated text-to-image generation using a diffusion model on supercomputer nodes.

  • Workshop · AI coding assistants

    AI Portfolio Workshop

    Workshop guide and prompt pack for building a personal portfolio site from organized context with an AI coding assistant.

03

Tools & Products

  • Featured · Shell · Jupyter · HPC

    Multi-Language Jupyter for HPC.

    Scripted installation of R, C, C++, Bash, and Julia kernels, extending interactive computing on the supercomputer beyond Python.

  • ARM64 Linux · Apptainer

    Home Server

    Bare-metal ARM64 Linux server (Fedora Asahi, Apptainer) mirroring the Grace architecture of the supported supercomputer, used for container and workflow testing.

  • Linux desktops · Browser-based

    Virtual Training Space

    Browser-based instructional computing environment providing each participant a prepared Linux desktop, with instructor visibility across sessions.

  • Verification · Audit trails

    Quiet Stack

    Tamper-evident record-keeping for AI interactions, producing independently verifiable receipts for each prompt and response.

WorkExperience

  • Deputy Director

    Morehouse Supercomputing Facility

    2026–Present

  • Faculty Fellow

    AUC Data Science Initiative

    2024–Present

  • Research Fellow

    University of Georgia

    2017–2024

Education

  • Ph.D., Bioinformatics

    University of Georgia

    2024

  • B.S., Mathematics

    Clark Atlanta University

    2017

PublishedResearch

Get in touch

Got a minute?

LET'S CONNECT.

I'm always interested in new opportunities, collaborations, and conversations. Whether you have a question or just want to say hello, feel free to reach out.

hope to hear from you →

© 2026 Ashley Scruse. All rights reserved.