Skip to content
View raulcmz's full-sized avatar

Block or report raulcmz

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
raulcmz/README.md

Hi, I'm Raúl 👋

Senior Platform / DevSecOps Engineer → AI Infrastructure & MLOps

9+ years designing secure platforms for mission-critical, regulated environments. Currently building the Kubernetes-native AI/MLOps layer on top of that platform engineering foundation — experiment tracking, model registries, distributed training, LLM serving, and the governance that lets ML teams move fast without breaking compliance.

📍 Spain · 🎓 MSc in Big Data & AI (2024–2026) · 🔐 Background in DevSecOps & software supply chain security


What I actually do

  • Platform engineering — internal Kubernetes platforms (Helm, ArgoCD/GitOps, Terraform, Ansible) that give product and ML teams self-service, standardized deployment paths.
  • Software supply chain security — SBOM generation, artifact signing (Cosign/Sigstore), SAST/DAST/SCA gates, and full traceability from source to running workload, applied across Rust, Python, Java, C/C++ and .NET stacks.
  • AI/MLOps infrastructure — Kubernetes-based platforms integrating MLflow, JupyterHub, MinIO, KubeRay and OpenMetadata for experimentation, model lifecycle management and metadata governance.
  • Observability — Prometheus/Grafana/Loki/Mimir stacks, plus LLM-specific metrics (TTFT, TPOT, token throughput) for inference workloads.

🔭 Featured projects

  • k8s-llmops-inference-platform — LLM inference gateway on Kubernetes (FastAPI, Ollama/vLLM adapters), with Prometheus metrics, an offline quality-eval harness, a baseline-vs-candidate promotion gate, and CI on every change. Platform-engineering patterns applied to LLM serving.
  • cybersec-mlops-pipeline — Master's thesis project: end-to-end MLOps pipeline for network anomaly detection (Isolation Forest), orchestrated with Prefect, versioned in MinIO, audited in PostgreSQL, running as Kubernetes Jobs on RKE2.
  • mlflow-k8s-ml-pipeline — End-to-end MLOps platform (MLflow, MinIO, FastAPI, Docker, Kubernetes) for training, registering and serving a churn-prediction model, with a stable champion alias for serving.
  • llm-prompt-evals — Prompt engineering and evaluation for real-world healthcare and finance use cases, focused on small local models, structured outputs and reproducible benchmarks.
  • ai-platform-notes — Public, versioned knowledge base (MkDocs) on ML infrastructure, MLOps, data engineering and DevSecOps-for-ML — written while building the projects above.

🌱 Open source contributions

  • kellnr/kellnr — merged fix for an OAuth2 callback URL bug (origin.path normalization) that broke OIDC login when deploying behind an ingress with a root path.

🧰 Core stack

  • Containers & Platform: Kubernetes Docker Helm ArgoCD GitOps
  • IaC & Automation: Terraform Ansible GitLab CI/CD Jenkins Python Bash
  • Supply chain security: SAST DAST SCA SBOM Cosign SonarQube Snyk OWASP ZAP Harbor Vault
  • Observability: Prometheus Grafana Loki Mimir
  • AI Infra / MLOps: MLflow Prefect JupyterHub MinIO KubeRay OpenMetadata FastAPI PostgreSQL

📜 Certifications

AWS Certified Solutions Architect · KCNA · CompTIA Security+ · CompTIA CySA+ · eCPPT · eWPT · eJPTv2 · Microsoft AI-900

📈 Currently

  • Completed an MSc in Big Data & AI (2024–2026) while shipping the projects above in production-style patterns (tests, CI, promotion gates).
  • Also studying Psychology part-time — it feeds directly into how I think about human-in-the-loop evaluation, model failure modes and building AI systems people actually trust.

📫 Reach me

LinkedIn

Pinned Loading

  1. k8s-llmops-inference-platform k8s-llmops-inference-platform Public

    Kubernetes-native LLM inference gateway (FastAPI + Ollama/vLLM) with Prometheus metrics, an offline quality-eval harness, and a baseline-vs-candidate promotion gate — CI on every change.

    Python

  2. mlflow-k8s-ml-pipeline mlflow-k8s-ml-pipeline Public

    End-to-end MLOps platform with MLflow, MinIO, FastAPI, Docker, and Kubernetes.

    Python

  3. llm-prompt-evals llm-prompt-evals Public

    Prompt engineering, evaluation and testing for real-world healthcare and finance use cases, with a focus on small local models, structured outputs and reproducible benchmarks.

  4. ai-platform-notes ai-platform-notes Public

    Public MkDocs knowledge base on AI platform engineering, MLOps, ML system design and DevSecOps for ML.

  5. cybersec-mlops-pipeline cybersec-mlops-pipeline Public

    📡 MLOps pipeline on-premises para detección de anomalías y correlación de incidentes en ciberseguridad mediante modelos no supervisados y reglas Sigma. 100 % OSS sobre Kubernetes.

    Jupyter Notebook