Roosa Yöruusu PRO

I am especially interested in applied AI engineering, agentic workflows, and building AI systems that are useful beyond demos. My current focus is on LLM-powered applications, RAG pipelines, evaluation, backend APIs, and multi-agent systems that can produce reliable, structured outputs. I am also interested in security-related AI from a defensive and purple-team perspective, such as AI-assisted detection engineering, MITRE ATT&CK mapping, automated triage, and tools that help analysts turn threat behavior into practical detection and response workflows. Recently, I have been working on projects such as AegisOps AI, a defensive “MITRE to Detection Copilot” built around multi-agent workflows, ROCm/vLLM inference, and structured security artifacts like Sigma rules, Splunk queries, response playbooks, and validation checks. My long-term goal is to become a strong AI engineer who can design, build, evaluate, and deploy end-to-end AI systems.