Writing
Notes on AI/ML & full-stack engineering
Deep dives on the systems I build — RAG pipelines, temporal graph models, explainable AI, and the infrastructure around them. Each post links its full technical report.
Graph MLKubernetes SecurityExplainable AI
Zero-Day: A Zero-Day Attack Detection & Mitigation Framework for Kubernetes
A six-layer framework that models Kubernetes container behaviour as a temporal graph, scores it with a TGNN + 3-model ensemble, explains every alert with XAI, and safely auto-contains threats through a Kubernetes operator.
Aug 2026·17 min read
RAGLLMsFull-Stack
Building a Production RAG Pipeline That Doesn't Hallucinate
How I took a naive retrieval-augmented setup from 61% to 92% answer faithfulness — chunking, hybrid retrieval, a reranker, and grounded generation with citations.
Jul 2026·2 min read