Writing
Blog
Insights on AI security, systems architecture, OSINT, and building intelligent platforms that solve real problems.
AI-Native Security Testing: Why Traditional Scanners Miss 80% of Modern Vulnerabilities
Traditional vulnerability scanners rely on static signatures and pattern matching that were designed for a simpler era. Modern applications built on microservices, SPAs, and serverless functions require a fundamentally different approach to security testing.
Building an OSINT Platform: Architecture Decisions Behind 300+ Data Source Integration
Integrating hundreds of disparate data sources into a coherent intelligence platform requires careful architecture decisions around normalization, graph modeling, and investigator privacy that most teams get wrong on the first attempt.
Multi-Agent AI Systems: Coordinating 16+ Specialized Agents for Complex Tasks
Single LLM calls fail at multi-step tasks requiring domain expertise across different specializations. The solution is multi-agent coordination, but the engineering challenges of making agents collaborate reliably are non-trivial.
LLM Security Vulnerabilities: When AI Models Become Attack Surfaces
Every organization deploying LLMs is introducing a new class of attack surfaces that traditional security tools cannot detect. From prompt injection to model extraction, the threat landscape for AI systems is expanding rapidly.
Systems Architecture in the AI Era: From Docker Swarm to Intelligent Infrastructure
AI workloads demand fundamentally different infrastructure than traditional web applications. GPU scheduling, model serving, polyglot persistence, and intelligent routing create new architectural patterns worth understanding deeply.