I build secure automation and local AI systems for infrastructure that cannot afford to fail.

Systems / Solutions Engineer working across data protection, Windows automation, detection engineering, and governed AI agents. My projects connect production infrastructure experience with practical Python, PowerShell, RAG, and security engineering.

           Security / Detection / Programming
               YARA rules · Powershell
                Python · C2 · OWASP
                        │
            ── BRIDGE ──┤── BRIDGE ──
        YARA pre-restore│  YARA → prompt injection
        malware-aware DR│  pattern defense for LLMs
                        │
  Backup & DR    ───────┼─────── Local LLM / RAG
  Veeam VBR             │        LangGraph · ChromaDB
  Network Security      │        BM25+RRF · Ollama
Systems Architecture    │        offline-first
                        │
            ── BRIDGE ──┤── BRIDGE ──
VBR data → AI or SIEM   │  topology = policy enforcement
    remediation agents  │  Agentic compliance & governance
                        │
              AI Safety | Agent Governance
               CyClaw   · Agentic AI drift detection
      SHA-256 integrity · MCP server
                        │
              Automation / Sysadmin (substrate)
       PowerShell· SCCM · HCI · VMware
       Azure · AWS  · Edge · Hyper-V

• Secure local AI: offline-first RAG, agent governance, policy-enforced workflows, prompt-injection defenses, and MCP-based tooling.
• Data protection and recovery: Veeam operations, malware-aware restore workflows, health-check automation, and resilient infrastructure lifecycle management.
• Detection engineering: YARA rules, ransomware indicators, suspicious infrastructure detection, and SIEM-ready output.
• Windows and platform automation: PowerShell, Python, SCCM, WinRM, VMware, APIs, and operational tooling designed for real environments rather than immaculate demo laptops.

The common thread: security policy should be enforced by system design, not merely requested in a prompt or buried in a runbook.

A local AI agent built around the principle that architecture should enforce policy. CyClaw uses LangGraph topology, hybrid retrieval, integrity checks, scoped tooling, and auditable execution to reduce reliance on prompt-only safeguards.

Python LangGraph FastAPI ChromaDB BM25 + RRF SQLite MCP Local LLMs

Why it matters: organizations need useful AI systems that can operate around sensitive data, constrained networks, and explicit governance requirements without quietly turning every control into a polite suggestion.

PowerShell and YARA tooling for detecting .onion infrastructure, cryptocurrency payment indicators, and command-and-control patterns in recovery data, with structured output suitable for investigation and SIEM workflows.

PowerShell YARA Veeam JSON Forensics Detection Engineering

Coordinates Veeam proxy availability with SCCM patching so maintenance can proceed without casually rebooting infrastructure beneath active backup jobs, a surprisingly popular human pastime.

PowerShell SCCM WinRM VMware Veeam

production constraint
      ↓
explicit threat / failure model
      ↓
architecture-enforced controls
      ↓
auditable automation
      ↓
operator-friendly outcome

I tend to optimize for:

• Local-first operation where privacy, cost, latency, or network isolation matter.
• Defense in depth rather than a single magical control with an impressive acronym.
• Dry-run modes, validation, logging, and rollback-aware workflows.
• Useful interfaces for operators, not just technically correct code that demands its own priesthood.
• Clear documentation and reproducibility so projects can be evaluated beyond screenshots and claims.

                 Security / Detection
            YARA · ransomware indicators · OWASP
                         │
       malware-aware DR  │  injection-pattern defense
                         │
Data Protection ─────────┼───────── Local AI / RAG
Veeam · recovery         │          LangGraph · hybrid retrieval
proxy lifecycle          │          local models · MCP
                         │
      operational data   │  topology-enforced governance
                         │
                 Automation Layer
          Python · PowerShell · SCCM · WinRM · APIs

• Expanding CyClaw's governed agentic coding and local-model support.
• Improving policy enforcement, evaluation, observability, and offline deployment paths.
• Turning hard-won infrastructure and recovery patterns into reusable security automation.