I build systems that think, ship & guard one engineer wearing the AI, full-stack, and security hats at once.
From orchestrating multi-agent LLM pipelines, to shipping enterprise Java & Flutter, to building a SIEM that hunts its own anomalies I like the messy seams between disciplines. That's where the interesting problems live.
Multi-agent orchestration, RAG & grounded reasoning that ships to production not to notebooks.
Enterprise Java backends, reactive web, cross-platform mobile the connective tissue from idea to user.
Detection engineering, SIEM pipelines & offensive tooling I build the defense and the attack to test it.
Planner/retriever agents on LangGraph + Spring Boot, grounded on pgvector & Redis, streamed over WebSocket. · LangGraph · pgvector · Redis · WebSocket
Self-hosted SIEM on Python · Kafka · Elasticsearch, with an Isolation Forest model mapping anomalies to MITRE ATT&CK.
Visual DAG engine React Flow + Spring Boot + Redis for resumable execution of OpenAI & Gemini agents.
Orchestrated Nmap + Metasploit pipelines with CVSS scoring and automated PDF pentest reports.
AST-based codebase indexing via Tree-sitter, wired to GitHub webhooks to auto-write PR descriptions.