AI Systems Engineer | AI R&D Engineer | Founder, Synthetic OS Labs
I design and build governed AI systems that combine language-model reasoning with memory, retrieval, deterministic computation, structured workflows, security controls, and meaningful human-review boundaries.
My primary work is Carter Synthetic OS — a public, runnable compound-AI research system integrating Carter, Synthetic OS, the Engineering Assistance System, the Synthetic Ideation System, and the Carter Sensory Console.
Carter Synthetic OS v0.1.0 is now publicly available as the Initial Public Research Release.
- Canonical repository
- v0.1.0 GitHub Release
- Architecture documentation
- Release notes
- Live Carter deployment
Carter Synthetic OS demonstrates how probabilistic language-model planning and generation can be combined with deterministic validation, computation, governance, evidence production, and explicit human oversight.
The public deterministic mock experience runs without private data, paid APIs, or a network connection. Local and cloud language-model providers are optional.
| System | Role |
|---|---|
| Carter | Flagship governed synthetic agent and conversational intelligence coordinating the integrated runtime |
| Synthetic OS (SOS) | Core architecture for orchestration, memory and retrieval interfaces, governance, context assembly, provider boundaries, and operational reporting |
| Engineering Assistance System (EAS) | Engineering decision-support workflow combining LLM planning with deterministic MCM computation, unit and constraint validation, engineering packs, decision records, governance gates, and required human review |
| Synthetic Ideation System (SIS) | Structured scientific ideation and invention workflow combining model-assisted exploration, feasibility analysis, evaluation interfaces, and governed hypothesis generation |
| Carter Sensory Console (CSC) | Session-isolated sensory workflow with explicit media controls, transcript handling, optional transcription and interpretation, text-to-speech integration, and local-only camera preview |
The July 31, 2026 release was verified against its exact tagged commit and packaged artifacts:
- 231 offline tests passed
- Python 3.11, 3.12, and 3.13 checks passed
- Clean Windows installed-wheel verification passed
- Ruff, Bandit, detect-secrets, dependency auditing, license inventory, and evidence reproduction passed
- CodeQL analysis completed with zero results for Python, JavaScript/TypeScript, and GitHub Actions
- All 18 public engineering-pack files were packaged and verified
- Source archive and wheel were independently inspected and published with SHA-256 hashes
- Released under the GNU Affero General Public License v3.0 only
These results apply to the exact v0.1.0 release snapshot and are not a claim of production certification or independent professional validation.
My work focuses on AI systems that are:
- Governed rather than purely reactive
- Structured around memory, retrieval, validation, and review
- Designed for engineering, scientific, operational, and reasoning-intensive workflows
- Capable of integrating both local and cloud-based language models
- Able to separate probabilistic model output from deterministic computation and validation
- Built with explicit security, authorization, privacy, and human-oversight boundaries
- Tested as complete systems rather than presented only as architectural concepts
The canonical public implementation of Carter, SOS, EAS, SIS, and CSC. It includes runnable Python source, a Flask interface, tests, synthetic evidence, engineering packs, governance boundaries, CI/security workflows, and extensive technical documentation.
A governed AI operator layer for human-supervised ArduPilot and Mission Planner workflows, including mission-state reasoning, bounded waypoint retargeting, and Sentinel Mode fire-detection concepts.
Focused public documentation and development history for EAS. The maintained runnable public EAS implementation is now integrated into the canonical Carter Synthetic OS repository.
Focused public documentation and development history for SIS. The maintained runnable public SIS implementation is now integrated into the canonical Carter Synthetic OS repository.
Earlier public architecture documentation for Synthetic OS and Carter. The canonical runnable public implementation is now maintained in Carter Synthetic OS.
- Governed and agentic AI architecture
- Compound-AI systems
- LLM orchestration and controlled prompt workflows
- Retrieval-augmented generation
- Long-term and working-memory systems
- Deterministic computation and validation
- Engineering decision-support systems
- Scientific ideation workflows
- Human-review gates and uncertainty boundaries
- AI application security and access control
- Local and cloud model integration
- Robotics and human-supervised AI operators
- Release engineering, CI, security scanning, and artifact verification
- Languages: Python, JavaScript, HTML, CSS, SQL
- Backend: Flask, Server-Sent Events, REST APIs, structured JSON workflows
- Data and memory: SQLite, retrieval pipelines, vector-memory interfaces
- Local AI: Ollama and local language models
- Cloud AI: OpenAI, Anthropic Claude, and Google Gemini APIs
- Voice and sensory systems: ElevenLabs, browser audio capture, transcription workflows, and local camera boundaries
- Testing and quality: pytest, Ruff, Bandit, CodeQL, detect-secrets, pip-audit, and GitHub Actions
- Development: Git, GitHub, virtual environments, package builds, wheels, and source distributions
The canonical Carter Synthetic OS repository contains a runnable public research implementation, including source code, UI components, tests, synthetic evidence, engineering packs, security automation, and technical documentation.
It intentionally excludes private prompts, credentials, private memories and conversations, operational data, production authentication, private infrastructure, cloned voice assets, and security-sensitive deployment details.
The public runtime demonstrates the architecture and integrated workflows without claiming to be behaviorally identical to the maintained private Carter deployment.
EAS outputs require qualified engineering review. SIS outputs require independent technical, safety, prior-art, patent, and experimental validation. CSC media capabilities require explicit activation and are not presented as identity, consent, or production-surveillance mechanisms.
I am a Computer Science graduate, U.S. Navy veteran, AI systems engineer, and founder of Synthetic OS Labs.
My background combines software development, electronics, systems thinking, engineering-oriented problem solving, AI research and development, and practical experience designing integrated generative-AI applications.
I am pursuing opportunities in:
- AI Engineering
- AI R&D
- Agentic and compound-AI development
- Generative-AI application engineering
- AI-focused software engineering
- Field and solutions engineering
I am particularly interested in roles where I can translate complex business, engineering, or scientific problems into governed, testable, and practical AI systems.
The maintained Carter deployment is authentication-protected. Guided demonstrations or temporary role-scoped access to Carter, EAS, and SIS may be available to prospective employers, researchers, collaborators, and pilot partners upon request.