Career architecture
Experience
Professional work history, technical leadership, and the systems built along the way.
Senior Staff AI/ML Engineer & Architect
Oct 2022 — Present
- Senior Staff AI/ML Engineer & ArchitectApr 2026 — Present
- Staff AI/ML Engineer | Technical LeadOct 2022 — Apr 2026
Lead AI/ML architecture and technical strategy for agentic AI systems across classified and unclassified defense programs. Guide multiple engineering teams across 8–10 concurrent efforts and drive organization-wide AI enablement for 500+ software and cyber engineers, with hands-on contributions spanning context engineering, multi-agent runtimes, MCP services, evals, guardrails, shared MLOps infrastructure, and distributed data platforms.
Achievements & technology stack
- Lead AI/ML architecture and technical strategy for agentic AI systems across unclassified and classified programs, contributing hands-on alongside AI, software, ML, and data engineering teams as systems move from design into production
- Serve on the engineering leadership team coordinating technical direction and architectural consistency across 8–10 concurrent AI/ML efforts. Help build multiple teams through ongoing technical interviews and hiring
- Serve as a senior lead for MFC's AI Enablement initiative, driving organization-wide adoption of AI tooling and agentic engineering practices among a workforce of 500+ software and cyber engineers. Develop and teach agentic engineering curricula, lead office hours, and mentor 10+ engineers across multiple teams
- Design and tune context-engineering systems for production agents, covering retrieval pipelines, summarization and compression, working and persistent memory, and context-window management to deliver relevant information throughout multi-step workflows
- Architect the runtime harness powering production agents, including the core agent loop, tool orchestration, state management, planning and stopping logic, error recovery, and multi-agent routing, delegation, and shared-state coordination
- Define and build the action surface for production agents by integrating internal APIs, databases, ticketing systems, and program-specific platforms through custom MCP servers with model-legible, safely scoped tools. Deliver multiple production agents and MCP services across programs using Pydantic AI and LangChain/LangGraph
- Build evaluation and observability infrastructure for production agents, including evals, tracing, logging, and monitoring. Champion evals-driven development and first-class observability from project inception
- Architect guardrails, security controls, and access management for production agents, including input and output filtering, prompt-injection defenses, permissioning, and scoped identity and authorization. Enforce each user's existing access rights and address security requirements that often determine deployment approval
- Architected a shared MLOps framework supporting several programs across classified and unclassified environments, spanning training, deployment, reproducibility, and data lineage. Cut production deployment time by roughly 85%, from about two weeks to under two days, and standardized releases across 10+ models and pipelines. Continue supporting MLOps and DataOps teams as a subject-matter expert through technical guidance and design reviews
- Diagnosed training bottlenecks in PyTorch-based deep learning models for anomaly detection using Torch Profiler and NVIDIA Nsight. Refactored the data loader and rearchitected key layers, cutting training time by over 94% on average, from at least 24 hours to under 90 minutes
- Built distributed data pipelines with Apache Airflow and Dask using functional programming patterns, processing 100–200 GB per run across 15+ data sources and tens of terabytes over time while roughly tripling throughput for ML and analytics workflows
- Spearheaded the architecture of an Apache Arrow-native data lakehouse spanning Amazon S3 and PostgreSQL, supporting 50+ TB of time-series data and metadata with sub-second analytical queries. Integrated an ML feature store serving 10+ production models
- Redesigned and normalized a core PostgreSQL schema, cutting query response times by more than 95%, from about 10 seconds to under 500 milliseconds, and tripling supported query and transaction throughput
- Deployed and operate a secure Kubernetes cluster in AWS GovCloud that supports 20+ ML and data visualization workloads with automated scaling and 99.9% application availability in a compliance-controlled environment
ML Engineer | Software Engineer | Data Engineer | Owner
Feb 2016 — Oct 2022
Owner-operated ML R&D practice (sister brand of Zia Technology Group) focused on ML-based algorithmic trading research and development, with selective external ML consulting.
Achievements & technology stack
- Founded and directed an independent ML R&D practice, defining the end-to-end technology strategy, roadmap, and agile delivery process across data, modeling, and algorithmic trading systems
- Engineered three open-source Python frameworks spanning data preparation through training, evaluation, and serving: Photon (deep learning lifecycle), Maxwell (market-data preprocessing and ML pipelines), and Dyson (financial modeling and algorithms)
- Designed deep learning and probabilistic models (CNNs, RNNs, Transformers, Bayesian neural networks, deep ensembles) powering algorithmic trading systems across equities, options, futures, and crypto markets
- Built GPU-accelerated (CUDA) data pipelines ingesting 20+ years of trade and order-book market data at multiple time resolutions, processing hundreds of millions of data points with Dask, RAPIDS, and Apache Arrow/Parquet
- Standardized research and production codebases on TensorFlow, PyTorch, Keras, NumPy, and Pandas, eliminating duplicated implementations and streamlining the path from experimentation to live trading
- Delivered external ML consulting, including a transformer- and autoencoder-based prediction system that improved a sports-analytics client's predictive accuracy by more than 65%
Data Engineer | Software Engineer | IT Consultant | Owner
Mar 2010 — Aug 2020
External data and software consulting brand of the same owner-operator practice as Applied Theta, working with client senior leadership to shape data-management technology strategies as consultant, architect, and lead developer for businesses across industries.
Achievements & technology stack
- Architected data integrations and ETL pipelines connecting 25+ business applications and data sources, including MS SQL Server, Salesforce API, and QuickBooks Enterprise. These systems automated reporting workflows and eliminated hundreds of hours of manual data handling annually
- Deployed and administered 40+ SQL/NoSQL databases with custom schema designs supporting client applications, data warehouses, and reporting systems
- Delivered full-stack web applications and reporting systems, including a client web portal that unified four backend systems in real time via Node.js and WebSockets for hundreds of business users
- Designed and implemented highly available cloud and hybrid co-lo/cloud virtual infrastructures for client environments, sustaining 99.9% uptime
Software Developer | DBA | IT Consultant | Owner
Jan 2007 — Mar 2010
Owner-operated consulting practice collaborating with client senior leadership to align technology with business needs while serving as consultant, DBA, and lead developer on critical IT infrastructure, data-integration, and web development projects.
Achievements & technology stack
- Administered 30+ MS SQL Server instances and databases powering custom and third-party business applications across client environments
- Designed and developed complex data integrations connecting multiple business applications and back-end data sources across client environments
- Developed web-based interfaces providing business users self-service access to back-end data sources, reducing manual data requests
- Led a team of 4 network engineers and developers delivering client infrastructure and software projects
Network Engineer | Software Developer | DBA | Owner
Dec 2001 — Jan 2007
Full-stack IT consulting practice supporting networks, servers, and applications for small and mid-sized businesses across a range of industries.
Achievements & technology stack
- Led a team of 15 technicians and network engineers supporting 200+ client networks and applications, improving on-time delivery and service reliability across engagements
- Deployed and supported hundreds of client networks, servers, and MS SQL Server database environments for custom and third-party applications
- Built dynamic, database-driven websites and web applications for small-business clients