cv

BA

Middlebury College

Physics and Mathematics

• Physics thesis: Structural Analysis of Potential Dual Quasar Pairs
• Mathematics thesis: An Introduction to Large Deviation Theory

Postbaccalaureate Researcher

National Radio Astronomy Observatory

Advisor: Adele Plunkett

• Compiling lists of nearby (<500 pc) Class 0/I protostars and Herbig–Haro objects from published catalogs
• Building a tested Python pipeline that harvests ALMA, MAST, NOIRLab, and other archive holdings and tags spectral-line coverage per target
• Identifying targets with multi-epoch observations suited to jet kinematics and proper motions

Research Intern

NSF-Simons AI Institute for Cosmic Origins (CosmicAI), National Radio Astronomy Observatory

• Built MANNA, an open-source MCP server that lets AI assistants query astronomical data archives via IVOA standards; released on PyPI and the MCP Registry
• Assessed MCP performance across multiple LLMs with a custom evaluation framework; with MANNA, models ran up to 8.1× faster at up to 6.7× lower cost
• Deployed open-weight LLMs on NOIRLab Astro Data Lab GPUs and integrated the Astro Data Lab JupyterHub with jupyter-ai

Lead Developer, STIPS

University of California Observatories, Lick Observatory

• Lead developer of STIPS, a package adapting the LSST Science Pipelines to 1-m class telescopes
• Calibrated Landolt standards to ≤12 mmag in BVRI; astrometry matches Gaia DR3 to 14–48 mas
• Reproduced published photometry of SNe 2023ixf and 2020wnt and recovered the HD 189733 b transit

Cataclysmic Variables Research Project

City College of San Francisco

• Observed cataclysmic variables, including AY Psc, in BVRI on the Lick 1-m Nickel Telescope
• Wrote a Python reduction pipeline: bias, flat-field, astrometric, and photometric calibration

AI/ML Engineer

Accenture Federal Services

• Led development for two small teams building production GenAI services
• Built over a dozen Python packages with unit, integration, and functional tests and CI/CD

Data Scientist

Accenture Federal Services

• Trained survival models on >10M UNOS records (Apache Spark) to predict kidney-waitlist outcomes
• Built ML models on NASA satellite imagery to predict wildfire, flood, and landslide risk

Research Assistant

Department of Physics, Middlebury College

• Cross-matched >1M known quasars with archival HST imaging; built a pipeline to reduce MAST data
• Ran surface-brightness contour analysis on thousands of images, identifying 21 dual-quasar candidates

Perityleae B.G.Baldwin (Asteroideae): A catalogue of natural history, taxonomy, and morphology of the rock daisy tribe

Capitulum, 4(2), 31

Lichter-Marck, I., Powell, A. M., Wogan, K., Gause, D., et al.

STIPS: Bringing the LSST Science Pipelines to 1-Meter-Class Telescopes

In preparation for PASP

Gause, D. & Westfall, K.

MANNA: an MCP Architecture for NOIRLab, NRAO, and additional Archives

In preparation for AJ

Gause, D., Nikutta, R., Mason, B., Plunkett, A., et al.

MANNA: MCP Architecture for NRAO and NOIRLab Archives

Poster, CosmicAI Cosmic Horizons Conference, Charlottesville, VA

Gause, D., Telkamp, Z., Plunkett, A., Mason, B., Nikutta, R., Juneau, S., Zacharia Anil, A., Loomis, R., Offner, S., Durrett, G., Torrey, P., & Murphy, E.

The Small Telescope Image Processing Suite (STIPS): Bringing the LSST Science Pipelines to 1-m Telescopes

Poster, 248th Meeting of the American Astronomical Society, Pasadena, CA

Gause, D. & Westfall, K.

On the Hunt for Dual Quasars

Keck Northeast Astronomy Consortium Symposium, Vassar College

Gause, D.

Searching for DQSOs in Archival Hubble Data

Poster, Middlebury Summer Research Poster Session

Gause, D.

Service Learner Volunteer

City College of San Francisco

• Helped plan and host monthly public sky viewing nights and planetarium shows

Scientific Computing Department Volunteer

California Academy of Sciences

• Built Python packages for herbarium data: DBSCAN clustering of specimens into collecting expeditions, and GBIF distribution maps for the rock daisy tribe (coauthor, Capitulum 2026)
• Parallelized a citation-matching app (40% faster); set up Apptainer under Slurm on the HPC cluster
• Trained models on EnMAP hyperspectral imagery to predict where tree species are absent