Computing · Data Science · Astrophysics

Hi, I’m Malia Barker.

I’m a PhD student in Computing (Data Science) at Boise State University working on reliable, transparent reasoning for large language models and computational tools for scientific discovery.

View my research See projects

Currently

  • 🎓 PhD student in Computing (Data Science), Boise State University
  • 🧠 Research with Dr. Edoardo Serra on improving reasoning in large language models.
  • ⚖️ Focus on reliability, transparency, and ethical deployment of AI in cross-disciplinary research.
Get in touch →

About

I am a PhD student in Computing with a focus on Data Science at Boise State University. My work centers on designing and evaluating methods that make LLM reasoning more reliable and transparent, with attention to how these systems are used in scientific and cross-disciplinary settings.

My background spans computer science, data science, and software engineering, with experience in building tools, exploring data, and communicating technical results to broad audiences.

Previously at Boise State University, I worked in astrophysics with Dr. Brian Jackson, developing software to analyze ultra-hot Jupiters and support high-precision orbital characterization.

Areas of interest

  • • LLM reasoning & evaluation
  • • Procedure-based reasoning and tool use
  • • Responsible and interpretable AI
  • • Exoplanets & ultra-hot Jupiters
  • • Scientific software & open-source tools

Links

Research

Reasoning in Large Language Models

With Dr. Edoardo Serra · Computing, Boise State University

I design and study procedure-based methods that structure how LLMs reason through complex problems, aiming to make their behavior more stable, verifiable, and interpretable.

  • Evaluation of structured vs. free-form reasoning
  • Reliability under stress tests and distribution shifts
  • Human-interpretable reasoning traces and error surfacing
  • Ethical considerations for deploying LLMs in research workflows

Exoplanets & Ultra-hot Jupiters

With Dr. Brian Jackson · Physics, Boise State University

I previously worked on modeling and observing ultra-hot Jupiters—gas giants on extremely close-in orbits—using software tools to support observation planning and transit timing analysis.

  • Transit and occultation timing for tidal decay searches
  • Software tooling for observation planning and data collection
  • Collaborative, open-source workflows for exoplanet science

Selected Projects

View all on GitHub →

Procedure-based LLM Reasoning Toolkit

A Python toolkit for defining, evolving, and evaluating structured reasoning procedures for LLMs, with support for scoring, validation, and comparative experiments.

The EvoProc Package →

Ultra-hot Jupiter Timing Analysis

Tools for combining transit and occultation observations to search for orbital decay in short-period exoplanet systems.

The Susie Package →

Data Storytelling & Public-Facing Reports

Projects combining tool-building with analysis of real-world datasets and communication for non-technical audiences.

Blog Posts →

Contact

I’m open to collaborations on research, open-source tools, and cross-disciplinary projects that bridge computing, data science, and science communication.

Quick facts

  • Based in Boise, Idaho.
  • PhD in Computing (Data Science) in progress.
  • Background in computer science, data science, and astrophysics.

For speaking, collaborations, or mentoring, email is the best way to reach me. I’m happy to share code, datasets, and workflows where possible.