Skip to the content.
Anthony Poole

Hi, I'm Anthony! πŸ‘‹

Software Engineer at Microsoft | Researcher in Applied Math

I enjoy working at the intersection of math, physics and computer science. Currently developing LLM-powered tools for code security at Microsoft while pursuing research in probabilistic modeling and chaotic systems.

Research & Publications

Research assistant at the Willcox Lab and with Ionut FarcaΘ™ at the Oden Institute, UT Austin (Oct 2023 β€” Present).

Convolution Operator Network for Forward and Inverse Problems (FI-Conv): Application to Plasma Turbulence Simulations

Xingzhuo Chen, Anthony Poole, Ionut Farcas, David Hatch, Ulisses Braga-Neto

Submitted to Journal of Computational Physics Β· arXiv

Short-Term Forecasting vs Long-Term Statistical Fidelity in Learned Turbulent Systems

Anthony Poole, Ionut Farcas et al.

Submitted to Journal of Computational Physics Β· email for pre-print

Probabilistic Scientific Machine Learning via Bayesian Model Reduction

Anthony Poole, Anirban Choudary, Karen Willcox

In progress Β· email for pre-print

Talks & Conferences

Microelectronics US β€” Speaker

April 22, 2026

Invited talk on digital twin technology and virtual prototyping. Try the live interactive demo: Ask Your Twin.

IPAM Workshop IV: Multi-Fidelity Methods for Robust Optimization and Real-Time Control of Fusion Processes β€” UCLA

May 18–22, 2026

Attendee.

Experience

Software Engineer - Microsoft

June 2025 β€” Present Β· Redmond, WA

Architected the AI agent harness and MCP for Microsoft's Code Analysis Platform, enabling automated vulnerability detection and fix generation. Built an A/B testing framework for evaluating agent configurations, improving fix quality by ~350%. Owned the platform's data pipeline using Databricks, Spark, and dbtβ€”cutting processing time by 500Γ—.

Software Engineer Intern - Microsoft

May 2024 β€” Aug 2024 Β· Redmond, WA

Designed a code understanding model enabling function-level search across Microsoft's codebase in multiple languages (92% on internal benchmarks). Built an advanced RAG system for security researchers, achieving 80% adoption and reducing average vulnerability research time by 60%.

Data Science Intern - Kodiak Robotics

Jan 2024 β€” May 2024 Β· Mountain View, CA

Built an Operational Design Domain model using data fusion, ensemble methods, and deep learning CV to classify in real time whether the AV was operating within its design envelope. Advanced the safety case through novel Bayesian statistical models informing engineering priorities.

Machine Learning Engineer Co-Op - AMD

Aug 2023 β€” Dec 2023 Β· Austin, TX

Designed an unsupervised image segmentation algorithm to quantify TIM delamination, reducing analysis error by >10%. Built a C++ Python package for geometric analytics, improving speed by >5,000Γ— and adopted by ~20 engineers.

Software Engineer Intern - Applied Research Laboratories

May 2023 β€” Aug 2023 Β· Austin, TX

Distilled U-Net and SAM segmentation models into smaller architectures for seafloor imaging, improving submarine battery life by 15% without sacrificing accuracy.

Education

University of Texas at Austin

B.S. in Physics | Minor in Computational Science

May 2025

Favorite Courses: Quantum information science, distributed systems, unconventional computation

Technical Skills

Core Competencies

Machine Learning Scientific Computing Security Analysis Data Science Inferential Statistics

Languages & Tools

Python C++ Java PyTorch JAX High Performance Computing

Research Areas

Turbulence Modeling Chaotic Systems Reduced Order Modeling Bayesian Methods SciML

Let's Connect

I'm always interested in discussing research collaborations, software, or really anything interesting!.