Confidential Defense-Technology Contractor · Autonomous Systems / UAV · Newport Beach, CA

• Contributed to a drone-based aerial mapping and reconnaissance system — processing real-time sensor feeds, spatial telemetry, and autonomous flight-path data for field operations.
• Developed components of a UAV mission-planning pipeline, including spatial mapping logic, obstacle-avoidance inputs, and real-time data integration for autonomous navigation.
• Applied reliability-first engineering standards in a production codebase where software accuracy directly impacts field operations.
• Operated under NDA and security protocols alongside senior engineers.

Topher Straus Fine Art · Fine Art Commerce · Denver, CO

• Appointed CFO at age 14 for a fine art business generating multi-million dollars annually — managed financial planning, budgeting, P&L reporting, and accounts payable/receivable for three years.
• Maintained financial models guiding pricing, collection valuations, exhibition ROI, and negotiations with high-net-worth collectors and institutional buyers.
• Architected a philanthropic giving program that distributed more than $250,000 across 15+ nonprofit organizations in arts, education, and community support.
• Oversaw operations: vendor contracts, event logistics, artwork shipping and insurance, and staff coordination for national exhibitions.

AI, Security & Systems Projects · Denver, CO

• Building Sectra, a voice-first AI assistant from scratch: a custom-trained wake word, a tuned speech pipeline, live camera object detection, and permission-scoped computer control running on local models.
• Built AI Agent Supervisor, an ML-engineering tool that monitors AI coding sessions for drift and prompt-injection, verified against real production session data up to 272MB with a test-first, zero-network-call test suite.
• Built a Brave/Chrome browser extension for security testing — a 31-tool catalog, live passive security scoring, and a native-messaging host bridging to a CLI security toolkit.
• Open-sourced PowerClaude, a personalization framework for AI-assisted development with 17 specialist configurations and a three-tier self-tuning system, published on GitHub under the MIT license.