ABOUT ME
cat dossier.txt
Iker Foronda Carrasco — Donostia/San Sebastián. Cloud & DevOps engineer and offensive security for AI applications: LLM red teaming, agentic OSINT and low-level traffic analysis.
Five prior years shipping product to production at an industrial scale-up: Full Stack (ML, WebRTC, PostgreSQL/Redis) and platform work on Kubernetes, Ansible, GitHub Actions and Terraform on Azure. Today, R&D in exposure management and offensive security at a cybersecurity product company.
The origin is academic: a master's thesis on adversarial campaigns against models, RAG systems and agents that became NORN, a public LLM red-teaming tool.
./cv --download
NOW
- Maintaining NORN, the LLM red-teaming tool that came out of my thesis, and looking for real cases to measure it against.
- Learning Rust, as a hobby.
- Open to talking about AI offensive security, cloud platform or product.
WORK AREAS
// cloud & platform
- Production application infrastructure on Kubernetes (AKS), with 50+ tenants managed.
- Infrastructure as code with Terraform, automation with Ansible and GitHub Actions deployments (CI/CD).
- Product observability and continuous delivery: from a day to 10 minutes per release (≈98% less time).
// AI offensive security
- Red teaming of LLMs and agentic systems: prompt injection (direct and indirect), jailbreak, RAG poisoning and tool abuse.
- Attack-success and error metrics —ASR, FAR/FRR, PSR@K, TDS, UAR— instead of anecdotes.
- Taxonomy mapped to OWASP Top 10 for LLM, MITRE ATLAS and NIST AI RMF; all of it published: NORN and the thesis.
// full stack & product
- Product shipped to production: backend and frontend, PostgreSQL and Redis.
- Machine Learning applied to industrial operations and live data visualisation with InfluxDB.
- Real-time communication with WebRTC in Node.js.
EXPERIENCE
April 2026 — Present
Senior Developer & DevOps
Cybersecurity product company — Donostia
- R&D in high-definition exposure management (defence posture, threat intelligence, external and internal exposure).
- Full-stack development of the product's multi-vendor aggregated data platform.
- Monitoring and observability of the product, plus DevOps and continuous delivery.
April 2024 — April 2026
Cloud Engineer
Industrial scale-up
- Production application infrastructure on Kubernetes (AKS), with 50+ tenants managed.
- Internal deployment platform: from a day to 10 minutes per release (≈98% less time).
- Automation with Ansible, automated deployments via GitHub Actions and infrastructure as code with Terraform on Azure.
October 2020 — April 2024
Full Stack Developer
Industrial scale-up
- Shipped product features to production, including advanced Machine Learning functionality for industrial operations.
- Backend and frontend development; PostgreSQL and Redis database management.
- Real-time communication with WebRTC in Node.js and real-time data visualisation leveraging InfluxDB.
March 2020 — May 2020
Full Stack Developer
Engineering and training consultancy
- Firmware and feature development for 3D printers.
EDUCATION & CERTIFICATIONS
// education
2024 — 2026
Master's in Cybersecurity and Privacy
UOC — Universitat Oberta de Catalunya
Thesis on LLM red teaming (distinction): the origin of NORN. read the thesis (PDF, 80 pp.) ▸
2016 — 2020
BSc in Computer Engineering
UPV/EHU — Donostia · Final degree project with distinction
// certifications
- GitOps at Scale — Codefresh · 2025
- GitOps Fundamentals — Codefresh · 2025
- Azure Data Fundamentals (DP-900) — Microsoft · 2022
- Azure Fundamentals (AZ-900) — Microsoft · 2022
- Azure AI Fundamentals (AI-900) — Microsoft · 2022
// languages
- Spanish — native/bilingual
- Basque — full professional
- English — full professional (C1)
SKILL MATRIX
self-assessment based on real production use · scale 0-10
currently learning: Rust
self-assessment based on real production use · scale 0-10