I'm Frangel — a cybersecurity engineer and full-stack developer working at the intersection of offensive security, resilient systems, and modern product engineering. From centralized SCADA deployments to CNN+LSTM phishing detection at 90.46% accuracy, I build things that have to hold up under pressure.
Click any card to dive in — full case study, gallery, features, and the engineering decisions behind each one.
Four overlapping practices. The interesting work happens where they meet — a SCADA rollout that needs a threat model, a phishing detector that needs to ship as a product.
Threat modeling, secure architecture, and hardening across web, mobile, and operational technology. From CIS benchmark automation to honeypot design.
Real-time monitoring of industrial fleets, secure protocol integration, and active/passive discovery for Modbus, S7Comm, and DNP3 environments.
End-to-end product development with a security mindset — from data modeling and APIs to polished frontends. TypeScript, Python, Flutter, and the occasional Go.
Hybrid CNN+LSTM models for phishing detection, NLP-driven threat analysis, and computer-vision pipelines with YOLOv8 and ONNX optimization.
The full index — every project from my GitHub and previous case studies. Click any row to open the full details.
Long-form thinking on the duality of AI in cybersecurity, automation tradecraft, and the gaps between what vendors sell and what actually keeps systems safe.
Who I am, what I care about, and the principles I try to bring to every engagement.
I'm Frangel — a recently graduated cybersecurity engineer and full-stack developer. My work sits at the intersection of offensive security research, defensive architecture, and product engineering. I build systems that have to be trusted rather than just liked: industrial control platforms, threat intelligence tooling, malware analysis environments, and the occasional consumer app where the engineering bar still has to be high.
I've shipped a centralized SCADA platform integrating industrial machines across multiple plants with secure Modbus/TCP and S7Comm protocols, an open-source ICS/OT scanner, a Windows honeypot framework, and a CNN+LSTM phishing detection model that hit 90.46% accuracy on PhishTank + Cisco Umbrella as my university thesis. I care about the whole lifecycle — from threat model to deploy to audit log — and I'd rather ship something small and observable than something large and opaque.
When I'm not engineering, I write about the duality of AI in cybersecurity, OSINT tradecraft, and automation. I'm currently available for select engagements where security and product quality are both non-negotiable.
Every system gets a threat model before a single line of code. If you can't name the adversary, you can't defend against them.
Layered controls are table stakes. What matters is being able to see, in real time, when one of them fails.
A small system you can reason about beats a large one you can't. Logs, metrics, and traces are part of the product.
Security that gets in the way gets bypassed. The right answer is controls people actually want to use.
I take on a small number of engagements per year — security architecture, ICS/OT hardening, malware analysis tooling, and full-stack product work where the security bar is non-negotiable.