I am a Senior AI Researcher bridging the gap between complex physical systems and reliable, trustworthy artificial intelligence. I specialize in taking complex research from rapid Proofs of Concept (PoC) to scalable, regulatory-compliant industrial solutions.
Currently, I lead the technical foundation of an internal AI Testing Lab initiative at DNV, focusing on:
- Agentic Systems: Architecting multi-agent systems using ReAct-style reasoning and action loops (LangChain, Google ADK).
- Enterprise AI & RAG: Designing robust Retrieval-Augmented Generation (RAG) pipelines via Azure AI Search and Vertex AI.
- AI Assurance & Safety: Developing comprehensive evaluation frameworks (e.g., DeepEval), execution pipelines, and AI-aware risk methodologies (CyberNEMO).
- Regulatory Alignment: Ensuring AI systems meet the stringent standards of the EU AI Act.
- Cyber-Physical Security: Secure co-simulation, adversarial testing, and malicious FMU development for digital twins.
- Core: Multi-Agent Orchestration, LLM Fine-tuning, Prompt-flow Orchestration, Generative AI.
- Assurance: DeepEval, Anomaly Detection, Uncertainty Quantification, Conformal Prediction.
- Core: Linux Power User, High-Performance Computing (HPC) basics, FastAPI backends.
- Cloud & DevOps: Azure AI Studio, Azure DevOps, GitLab, End-to-end CI/CD deployments.
- Senior AI Researcher | DNV, Norway (Apr 2026 - Present)
- Technical lead for an internal AI Testing Lab, architecting agentic execution pipelines and use-case-driven testing frameworks.
- AI Researcher / Global Research Trainee | DNV (Feb 2023 - Mar 2026)
- Developed ReAct-based multi-agent architectures (Assurance Copilot), an EU AI Act demonstrator, and conducted adversarial co-simulation testing.
- Doctoral Research Fellow | University of Oslo (2018 - 2022)
- Investigated high-resolution solar phenomena using ALMA data and physics-informed computational simulations.
- Visiting Research Scholar | UC Berkeley, Space Sciences Lab, USA (2018)
- Researched ground station design for the CURIE CubeSat and led image processing pipelines for the 2017 total solar eclipse.
| Degree | Institution | Year |
|---|---|---|
| Ph.D. in Astrophysics | University of Oslo | 2022 |
| M.Sc. in Astronomy & Astrophysics | Universidad Nacional de Colombia | 2018 |
| B.Sc. in Physics | Universidad Nacional de Colombia | 2015 |
Contact: juancamilo.guevaragomez@gmail.com
Location: Oslo, Norway 🇳🇴
💡 Note: Most of my recent enterprise architecture and agentic system developments are hosted on internal DNV Azure DevOps and enterprise GitHub repositories.