Data Inputs
NOAA NWM, ERA5, USGS observations
M.S. Artificial Intelligence · UT Austin
M.S. AI student at UT Austin
I build end-to-end ML systems: model architecture, temporal validation, reproducible pipelines, and production APIs. My thesis achieved 48% streamflow error reduction using hybrid deep-learning. TS/SCI cleared and available now.
Available Now · Full-Time or Summer 2026 Internship
Featured Project
Hydrologic forecasting designed for operational reliability
Thesis Project
HYDRA
Hybrid Deep-learning for Residual Analysis
National water models often struggle with local precision. During events like Hurricane Helene, NWM forecasts significantly underestimated peak flows in complex terrain, leaving communities with uncertain flood warnings. Achieved up to 48% improvement in predicting streamflow across unregulated Appalachian watersheds. Live at hydramodel.ai. Manuscript in preparation for Water Resources Research (AGU).
- Up to 48% improvement in predictions across unregulated Appalachian watersheds.
- Consistent improvement across varying forecast horizons and extreme storm events.
- Demonstrated stability under seasonal distribution shifts.
PyTorchTransformerHydraMLflowxarrayAWS Batch
Model Core
Transformer encoder with GRU residual head
Outcome
Up to 48% streamflow error reduction
About
Engineering rigor, research depth, and collaborative delivery
My background spans high-stakes operational environments, production software engineering, and applied ML research. Across all of them, I've learned that strong systems come from clear interfaces, disciplined validation, and good communication between people with different expertise.
My senior thesis achieved up to 48% streamflow error reduction in national water forecasts using hybrid deep-learning architectures. I'm especially effective where research, engineering, and domain expertise intersect. Reliability and clarity matter as much as raw performance.
Career Timeline
SWE Intern
May 2025 - August 2025USAA
- Designed and implemented GraphQL APIs using Java and Spring Boot to surface core customer data.
- Enhanced internal troubleshooting tools by integrating Non-Source-of-Record data.
- Built and refined JavaScript front-end components for data visualization.
- Collaborated in an Agile/Scrum team using Jira and Git.
- Worked closely with product owners, backend engineers, and internal users to ensure APIs were usable, secure, and operationally reliable.
Executive Missions Aviator
August 2020 - April 2023United States Air Force
- Maintained passenger safety and schedule reliability for distinguished guests aboard Air Force 2.
- Coordinated across flight crew, security teams, and executive staff to meet exacting operational standards.
Managing logistics for executive missions under zero-margin-for-error conditions shaped how I think about reliability in every system I build.
How I Work
I ask a lot of questions early to avoid costly assumptions later
I value clear ownership and well-defined interfaces
I document decisions so teams can move faster over time
I'm comfortable bridging research ideas and production constraints
Additional Projects
Systems built to operate under real constraints
From regulated financial services to algorithmic trading, my work focuses on reliability, evaluation, and long-term behavior, not just model accuracy.
USAA Risk Services
Secure Production APIs for Financial Services
Problem
Internal teams needed a reliable, compliant way to access core risk data without navigating legacy complexity.
Why it matters
Built production APIs that partner teams across the organization now use daily. Worked closely with product owners, backend engineers, and internal stakeholders to make risk data accessible and self-serve.
Approach
- Designed and deployed strict GraphQL APIs using Java and Spring Boot.
- Enhanced internal observability tools by unifying disparate data sources.
- Built responsive React dashboards to give stakeholders visibility into risk metrics.
Stack
JavaSpring BootGraphQLReactTypeScriptPostgreSQL
Results
- Consolidated fragmented data access into a single self-serve API layer, eliminating recurring partner-team support loops.
- Introduced schema-validated deployment gates that caught contract regressions before production.
What I learned
- Clear interfaces (API contracts) allow teams to move fast safely.
- Effective production systems require close collaboration between engineers, product owners, and end users.
Active Development
Harmony
Algorithmic trading systems for disciplined decision-making
Problem
Financial markets are noisy, adversarial environments where predictive accuracy alone is insufficient. Sustainable trading systems require disciplined evaluation, strict risk controls, and infrastructure that behaves predictably under stress.
Why it matters
A modular trading system with independent forecasting, risk, and execution components. Each module can be evaluated, replaced, or improved independently.
Approach
- Design a forecasting and execution pipeline that survives regime shifts.
- Emphasize risk management over raw returns.
- Build components that can be evaluated independently and replaced safely.
- Treat trading as a systems engineering problem, not a single model.
Stack
PythonPyTorchpandasNumPyPostgreSQLDocker
Current Scope
- Market data ingestion and normalization pipeline.
- Forecasting models for short- and medium-horizon signals.
- Risk engine enforcing position sizing and drawdown limits.
- Backtesting framework with leakage-aware evaluation.
- Execution simulation via paper trading.
Design Principles
- No model is trusted without adversarial backtesting.
- Risk constraints override model confidence.
- Components must fail safely.
- Performance claims require statistically sound evaluation.
Contact
I'm looking for teams building production systems that matter
I'm seeking full-time AI/ML or software engineering roles on teams that value reliability, clarity, and execution. Open to industry or research environments.
Schedule a call
Pick a time and a Zoom link will be generated automatically.
2026 Mitchel Carson.