About Brian James Curry

Brian Curry is an enterprise AI and decision science leader, machine learning practitioner, published applied researcher, and open-source framework developer working at the intersection of machine learning, causal inference, MLOps, intelligent decision systems, knowledge engineering, and the economics of AI.

He is the founder of Vector1 Research, an independent applied AI lab developing and publishing original research across machine learning, causal inference, Bayesian modeling, complex systems, marketing economics, agentic AI, forecasting, cognitive architecture, MLOps, and AI-driven organizational transformation. His work focuses on problems where conventional analytics are insufficient: how models learn from complex systems, how intelligent systems make decisions, how those decisions should be evaluated, how causal effects can be separated from correlation, how machine-learning systems can be operationalized reliably at enterprise scale, and how AI changes the economics and operating models of firms.

Brian has spent more than 20 years across analytics, digital, product, marketing, and technology, including extensive people leadership and P&L responsibility and nearly a decade of hands-on machine learning, data science, experimentation, causal inference, applied AI, and ML engineering. His enterprise experience includes Vail Resorts, Tractor Supply Company, Koch Industries, Hallmark, Garmin, AT&T, and McClatchy, spanning Fortune 300 retail, hospitality, industrial markets, telecom, consumer brands, and digital marketplaces.

A major area of his enterprise work is machine learning and decision science: predictive modeling, customer understanding, segmentation, propensity and response modeling, personalization, next-best-action systems, experimentation, causal measurement, forecasting, optimization, and the governance required to scale automated decisions responsibly. His work connects model development to real operating decisions rather than treating machine learning as an isolated technical exercise.

Brian also has experience designing and working within MLOps and production machine-learning environments across Microsoft Azure, Amazon Web Services, and Google Cloud Platform. His work spans the machine-learning lifecycle from experimentation and reproducible model development through data and feature pipelines, model packaging, deployment, orchestration, versioning, CI/CD, model registry and artifact management, monitoring, retraining, and production governance. He has worked across cloud-native and enterprise data environments and understands how to translate locally developed models and research prototypes into scalable, observable, maintainable ML systems. His approach to MLOps emphasizes reproducibility, separation of training and inference concerns, automated testing, traceability, model and data quality, responsible deployment, and continuous evaluation of model performance after release.

This combination of research, ML engineering, cloud architecture, and decision science allows him to operate across the full AI lifecycle: identifying a high-value business problem, developing the modeling methodology, validating causal or predictive performance, designing the surrounding system architecture, operationalizing the model within enterprise cloud infrastructure, and establishing the monitoring and governance necessary to keep the system reliable over time.

His published research and technical writing spans machine learning, Bayesian marketing measurement, causal inference, optimal control, multi-agent systems, AI evaluation, knowledge graphs, forecasting, economic simulation, cognitive AI, ML systems, and the structural effects of automation on firms, labor, and capital. His formal research is also published through repositories including Zenodo.

Open-source work includes Papilon, a Python framework for machine learning, causal modeling, Bayesian MMM, simulation, economic systems, and intelligent decisioning; PyCausalSim, a simulation-driven causal discovery and evaluation framework; MeaningFlow, a semantic intelligence framework using embeddings, clustering, dimensionality reduction, and graph methods; and Memory-Node Encapsulation, a cognitive architecture for artificial episodic memory and causal reasoning in agent systems. Across this work, Brian emphasizes reproducibility, modular system design, evaluation, experimentation, and the progression from research code toward operational machine-learning systems.

In 2016, Brian founded KC AI Lab, originally the Kansas City Machine Learning Group. The organization grew to more than 1,500 AI and machine-learning practitioners and became one of the largest community-built AI organizations in the Midwest.

Brian grew up in a small farming town in Kansas and began working crop fields at twelve. He attended Kansas State University as a collegiate athlete and later trained Brazilian Jiu Jitsu. Outside of research and technology, he is a musician, composer, and painter. His creative work explores improvisation, generative systems, and human–AI collaboration. He is the father of two.

Research and open source: github.com/Bodhi8
Vector1 Research: vector1.ai
LinkedIn: Brian Curry