Data scientist and AI engineer with a PhD in Artificial Intelligence and 10+ years of industry experience delivering machine learning, NLP, optimization, and generative AI solutions across construction (EPC/AEC), energy, retail, and government verticals. Expert at translating complex operational and engineering data into production-grade systems and decision-ready insights. Practitioner of the full modern data science stack — from exploratory analysis and model development through cloud deployment, orchestration, and stakeholder-facing dashboards.
Atos zData
Delivered AI and data science solutions to complex business problems across EPC, AEC, energy, retail, defense, and government verticals. Led engagements spanning discovery, model development, and production deployment.
University of Maryland Baltimore County
Conducted NSF-funded research in geospatial analytics, scientific visualization, and collaborative science. Architected and devloped the GLOBE system. Mentored graduate and undergraduate students.
University of Maryland Baltimore County
Developed a natural language–to–sketch (TTS) system for the National Geospatial-Intelligence Agency enabling analysts to produce geospatial movement sketches from unstructured text inputs.
University of Maryland
Researched metacognition in autonomous systems. Developed Bayesian network–based ontologies for failure diagnosis and recovery to enable the next generation of robust, self-correcting AI agents.
FBB Bass Works — Self-employed
Owner-operated boutique workshop specializing in handcrafted custom electric bass guitars. Washington DC–Baltimore area.
Client names are anonymized; descriptions identify industry and scope.
For a global engineering, procurement, and construction company: engineered a fiber span prediction algorithm to streamline permitting; built an NLP-powered risk report dashboard; developed a Primavera P6 schedule alignment tool with what-if analytics; and led AI-assisted cost code unification across a diverse project portfolio.
Conducted stakeholder discovery and developed a proof-of-concept AI engineering assistant for a multinational architecture, engineering & construction firm as part of a digital transformation and AI modernization initiative.
Led a cross-functional team delivering a Monte Carlo–based intelligent scheduling assistant for a military training program. Replaced manual, error-prone planning with data-driven analytics providing bounded-certainty schedule forecasts under high operational uncertainty.
Modernized and orchestrated predictive modeling, resource management, and optimization pipelines for a multinational energy company, covering hydro, solar, wind, and fossil sources using neural and statistical time-series methods.
Contributed to a multidisciplinary team building an enterprise-grade virtual concierge for a state-of-the-art performing arts venue, using tool-calling LLMs and developing supporting analytics, monitoring, and BI tooling.
Developed an on-premises virtual assistant for a public university using open-source models and a RAG backend to provide students with support across a wide range of administrative and logistical topics.
Contributed to the development of a RAG-based digital avatar for a global company operating in the security and operational efficiency space.
Built analytics, AI training pipelines, and automation tools for a next-generation brick-and-mortar retail monitoring platform enabling real-time store intelligence at scale.
Assessed the AI readiness of a commercial fleet software company's data assets and scoped a portfolio of AI modernization projects. Produced recommendations to strengthen data infrastructure for an AI-enabled product roadmap.
Led the data science effort to unify surveys, product reviews, and social media for a major apparel retailer into a single source of truth, with BI and NLP tools to surface actionable customer insights.
Replaced a manual, spreadsheet-based size ratio process for a major athletic apparel brand with a modern, automated in-house application — reducing error rates and planning cycle time significantly.
Developed a dashboard and predictive analytics suite for an online music media platform to better understand customer behavior and surface up-and-coming content and tastemakers within the user base.
GLOBE: Analytics for Assessing Global Representativeness
5th International Conference on Computing for Geospatial Research and Application
Methods for formalizing and visualizing the representativeness of ecological study site collections and correcting sampling bias, applied to land change science.
The Metacognitive Loop and Reasoning about Anomalies
Metareasoning: Thinking about Thinking — MIT Press
An architecture for generalized metacognition in AI systems that monitors expectation violations, diagnoses failures, and plans recovery in an application-general manner.
Mining of Concurrent Text and Time Series
KDD Workshop on Text Mining
A novel approach to identifying news stories that influence financial market behavior by jointly mining text and time-series signals.
Learned Models for Continuous Planning
AI & Statistics
An approach to generating dynamical models of activity from real-world experience and applying them toward planning in a continuous state space.