Roosa · Finland · Open to junior/mid roles
Reliable AI workflows and secure Python backends.
I build multi-agent automation and backend systems with validation, testing, and deployment in mind.
Selected work
AI security automation
Automates the path from MITRE ATT&CK techniques to Sigma rules, SOC guidance, and validated output.
Built as a four-agent LangGraph pipeline with a Qwen verifier and a public Hugging Face demonstration. A vLLM inference path runs on AMD ROCm hardware.
Production delivery
Delivered a production system connecting work orders with pipe and material inventory.
Includes stock deduction, remnants, audit trails, authentication, health checks, handover documentation, and Linux deployment. Client details remain confidential; the repository is a sanitized demonstration.
GitHubEvaluated RAG
RAG-assisted Python security auditor with citations and reproducible evaluation.
Backed by 20,875 embeddings and a documented 15-case evaluation suite, with FastAPI, pgvector, pytest, and GitHub Actions CI/CD. Baseline results and failure analysis remain public in the repository.
GitHubHow I build
My cybersecurity background shapes how I approach AI and software: model output is validated, failure modes are made visible, and deployment and handover are part of the system—not afterthoughts.
LangGraph, RAG, local inference, structured outputs, and evaluation suites.
See AegisOpsFastAPI, PostgreSQL, REST integration, Docker, systemd, nginx, and Linux deployment.
See client deliverypytest, GitHub Actions, validation, authentication, audit trails, and operational handover.
See evaluated RAGAI engineering focus since January 2026
Building AI workflows, backend applications, and deployment-ready tools from system design through testing and documentation. Work includes multi-agent security automation, RAG evaluation, local-first AI, and confidential client delivery.
Worked independently in a regulated home-care setting where reliability, discretion, and clear communication with non-technical clients were essential.
Analyzed system architectures and networked environments to identify weaknesses and improve reliability. Built scripts and tooling for analysis workflows and system validation.
Built Python tools for data processing, monitoring, workflow automation, and system-level integration.
Log analysis, documentation, and security monitoring support.
Additional technical activity: independent security research from 2018 to 2026, including web-system analysis, proof-of-concepts, and public security tooling. Earlier roles and additional detail are available on LinkedIn.
Selected recent learning: AMD Developer Hackathon completion · Google AI · IBM Generative AI
Languages: Finnish (native) · English (professional) · Norwegian (elementary)
Professional recommendations
“Roosa’s knowledge of IT security and her skills are something to be amazed at. A hearty recommendation.”
Mikko Väisänen
Chief Information Security Officer, theFIRMA
Same-team colleague ·
“Roosa is a bright individual; she had no trouble at all at passing her courses with the highest grades. Also, she is a delight to work with.”
Vesa Slotte
Cisco Networking Academy Instructor
Former teacher, Turku University of Applied Sciences ·
Contact
Based in Finland and available for junior/mid opportunities. Email is the quickest way to reach me.
Messages sent to contact@ztothez.com are used only to respond to your enquiry.