Defense Program Analyst writing about autonomous systems, Counter-UAS, and why our acquisition process can't keep up with the drone threat.
I bridge two worlds: the structured discipline of federal defense contracting and the scrappy full-stack ownership of entrepreneurship.
Before I walked into my first Navy building, I spent 9 years running a German automotive diagnostics operation in San Diego — growing it from $600K to $1.2M, leading an 8-person team, and learning that every problem is a systems problem.
Then I brought that mentality to the DoD. At CNIC and Booz Allen Hamilton, I supported Navy autonomous vessel programs (PMS406), built $23M budget submissions through the PPBE process, deployed Counter-UAS capabilities, and coordinated with everyone from SURFDEVRON to NSA.
I'm not the person who makes the strategy at DARPA. I'm the person at the installation level who has to make the strategy work — and who writes the whitepaper when it doesn't.
Active DoD Secret
English, Spanish, Polish (native), Italian
San Diego, CA (Open to Remote)
Navy Programs (BAH) → GS-13 → Entrepreneur
The $500 drone vs. $3M interceptor cost exchange. Why directed energy can't come fast enough and why buckshot isn't the answer.
PPBE takes 2+ years. The threat evolves in months. Replicator promised thousands, delivered hundreds. The process is the problem.
USV/UUV programs, ORCA vs. Ghost Shark, fleet integration challenges. Why autonomous vessels need policy before propulsion.
CMMC compliance threatening 300K+ contractors. Deepfake voice attacks on defense leadership. The human factor in force protection.
3M+ DoD users on GenAI. IL5 vs. IL6 security levels. The gap between ChatGPT demos and classified network deployment.
Nobody gives you a manual when you walk into your first Navy building. The org chart is a lie. Acronyms are a test. Outside perspective is your superpower.
Defense analysis grounded in installation-level experience. No DARPA abstractions — real problems, real numbers.
Iran launched 541 drones at Gulf states. Defense systems stopped 506. But 35 got through — and one hit the Fifth Fleet's headquarters in Bahrain. After $1B+ in Red Sea interceptors, the math still doesn't work.
The US reverse-engineered Iran's Shahed-136 and used the clone against Tehran in Operation Epic Fury. But one copied drone doesn't make a strategy — Ukraine is building 7 million this year.
In Operation Epic Fury, cyber operators struck first — blinding Iran's communications before a single bomb dropped. Now 60+ hacktivist groups are retaliating, and CISA is at 38% capacity.
China spent 30 years building a kill chain to push carriers back 1,000 miles. What happens when the fleet sends 200 autonomous vessels worth $50K each instead?
In 2022, a senior leader told our team the backup plan for a small drone attack was buckshot. I spent the next week writing a whitepaper on why that would get people killed.
We spend billions building autonomous systems, then starve them of the sustainment funding they need. The "acquisition valley of death" isn't about buying — it's about keeping.
Boeing's ORCA: $885M and 8 years. Anduril's Ghost Shark: ~$100M and 3 years. Same mission, radically different acquisition models.
AI-generated voice cloning can now replicate a Flag Officer's voice from 3 seconds of audio. The next generation of social engineering attacks won't come through email.
I spent 9 years diagnosing German cars. Then I walked into a Navy building and nobody gave me a manual. The transition was humbling — and my outside perspective became my edge.
The DoD has 3M+ users on commercial GenAI tools. None of it works on classified networks. The gap between "use AI" and "use AI where it matters" is wider than anyone admits.
The Pentagon promised thousands of autonomous systems by August 2025. It's 2026. We got hundreds. If your doctor says lose 50 pounds and you lose 12, nobody calls that a win.
I didn't know what an N-code was. I didn't know "the Flag" meant an admiral. I didn't know a "P-halt" could kill your entire program in a single email.
What Ukraine built in months, our acquisition system can't deliver in years. That's not a technology problem. That's a process problem.