"You are unstoppable when determination is reinforced by resilience and diligence."
I'm Evangelos, but most people simply know me as "Angelo". I build AI systems that focus on solving hard problems like cancer research, agentic space tech, and ambient signals for proactive agents.
My professional career propelled forward as a technical controller and cybersecurity lead for the United States Marine Corps as part of a joint task force helping stabilize areas with cyber awareness and security practices. I've helped lead initiatives for AI/ML hardened workloads in the DoD, written Big 4 advisory white papers on AI/ML security for enterprise cloud, and innovated new methods for security and intelligence with WIRESHIELD and FACTION.
I understand the importance of pacing towards the forefront of bleeding edge to help solve the modern challenges of today and tomorrow. This is why I built the FACTION agent ecosystem with a Rust engine that can run on a local machine, in VS Code as an extension, and at the cloud's edge.
The speed at which adversarial technologies are leveraged is not slowing down. So my protector mindset inspired me to create WIRESHIELD to deliver a new design and method of detection systems to the benefit of privacy, safety, and awareness.
It's a brave new world, thrust upon us, and I'm here for it.
A few things I actually operate by. Specific, not slogans.
I operate by a protocol of extreme focus to understand the topic at hand from a macro and micro level of impact areas and expected outcomes. This is imperative for effective decision making without slowing down the velocity of productivity.
I stick by first principles in systems engineering logic when working with AI to prevent unnecessary complexity and simplify workflows.
Am a firm believer that a first impression is an incredible way to understand people if you pay close enough attention. It will help you understand nuance that carries forward.
Never be so foolish as to think you cannot learn something from anyone who walks in a room or provides feedback at any given time. Nearly everything around you is a data point and the generosity of others who offer perspective can provide incredible value depending on how you interpret it.
What's on the stack right now. Refreshed periodically.
Open questions I'm working through.
How do you make agentic systems coherent across sessions when the underlying model has no memory of you? This is a problem that many have approached with different types of solutions and many falling in the category of extended memory systems. Adding more agents to the equation shifts the paradigm. Who should know what, and when? And, what happens when information is stale and there are duplicative entries and collisions on data points? This is a concept FACTION helps resolve with SKYMAP context data mesh and its native agentic firewall. The results are nearly as exciting as what the future holds.
How will open models help augment demand for compute with frontier AI? Is WebGPU the lowest friction and best solution for the everyday user? This is a fun one to mull over since FACTION runs local and frontier models, and at the edge via WebAssembly. Packaging it with a high-performing small local model as a progressive web app is an interesting concept I've been thinking through lately.
Direct lines. I read everything that comes in.