Profile

About Joseph Adeleke, MSc

"To be, is to do." — Versatile engineer specializing in Multi-Cloud DevSecOps and MLOps. Focuses on secure, scalable cloud architectures and AI-driven workflows, with a strong enthusiasm for automation and bridging the gap between development and operations.

          What I bring
          Currently owning end-to-end DevSecOps and MLOps delivery for MTN Ghana and enterprise clients, managing 150+ CI/CD pipelines.
Expertise in building and managing GPU-enabled MLOps infrastructure on AKS using Ollama, Kubeflow, and MLflow for rapid LLM/ML deployment.
Deep experience managing 10 Kubernetes clusters across dev, UAT, and production environments, enforcing strict RBAC and zero-trust policies.
Proven integration of automated security gates (Trivy, SonarQube, Azure Key Vault) within CI/CD workflows at national telecom scale.

        

Reference MLOps Architecture

This is my standard approach to producing reliable, observable AI systems from raw data to end-user serving.

01. Data Sources

02. Ingestion & Validation (Great Expectations)

03. Feature Engineering (PySpark, Polars)

04. Train & Evaluate

05. Model Registry (MLflow, DVC)

06. CI/CD Build & Test (GitHub Actions)

07. Containerize & Deploy (Docker, Kubernetes, Azure)

08. Online Serving (REST/gRPC, Triton)

09. Monitoring & Logging (Grafana, Datadog)

10. Drift Detection & Retrain