Stephen Sequenzia
Senior Staff AI/ML Engineer & Architect
Across more than 20 years of hands-on engineering, technical leadership, and business ownership, the focus has remained consistent: building reliable systems, putting them into production, and delivering measurable results.
Selected signals
Evidence at operating scale
8–10
Concurrent AI/ML efforts
Architecture and technical direction across classified and unclassified programs.
500+
Engineers reached
Organization-wide agentic engineering enablement across the software and cyber workforce.
~85%
Deployment-time reduction
Reduced production deployment from about two weeks to under two days.
>94%
Average training-time reduction
Reduced PyTorch training from at least 24 hours to under 90 minutes.
Selected work
Systems built to operate
Agent Alchemy — Claude Code Toolkit
An ecosystem of 6 Claude Code plugin groups (28 skills, 16 agents), a real-time task-manager dashboard, and a VS Code extension
System stack
- Claude Code
- Markdown-as-code
- TypeScript
- Next.js
- React
- +4
Capo — Personal Agent Orchestrator
A long-lived AI orchestrator that ingests iMessage/Discord and delegates coding to Claude Code/Codex via restart-safe durable workflows
AMG — Agent Messaging Gateway
A messaging gateway exposing a four-tool MCP + REST surface over iMessage and Discord behind one normalized envelope
- 2022–nowSenior Staff AI/ML Engineer & ArchitectLockheed Martin
- 2016–2022ML EngineerApplied Theta
- 2010–2020Data EngineerZia Technology Group
- 2007–2010Software DeveloperThinksys
- 2001–2007Network EngineerBiziteks
Profile
The longer view
Across more than 20 years of hands-on engineering, technical leadership, and business ownership, the focus has remained consistent: building reliable systems, putting them into production, and delivering measurable results. At Lockheed Martin, that work now centers on leading AI/ML architecture and technical strategy for agentic AI systems across classified and unclassified defense programs, guiding multiple engineering teams across 8–10 concurrent efforts, and driving agentic engineering enablement for 500+ software and cyber engineers. The role remains deeply hands-on across context engineering, multi-agent runtimes, custom MCP services, evals, guardrails, MLOps, and distributed data platforms. Previous work includes founding and operating technology consulting and ML R&D businesses, as well as authoring open-source frameworks spanning the machine learning lifecycle, data engineering, and financial modeling. Across these efforts, the focus remains reliable systems and measurable outcomes, including an 85% reduction in ML deployment time and an average reduction of over 94% in deep learning training time.