TIM POMAVILLE
TIM POMAVILLE
I am a multidisciplinary engineer with a foundation in both Software and Chemical Engineering, and a background spanning the automotive, chemical, and aerospace industries. That industrial context shapes how I build data systems - I understand the processes the data describes, not just the pipelines that move it.
I work as a Platform & Data Engineer, designing and operating Kubernetes-native infrastructure for data-intensive workloads. My stack spans GitOps with ArgoCD, secrets management with Vault, SSO with Keycloak, and a full ML/data platform including Spark, Airflow, MLflow, Trino, JupyterHub, and Metabase - all deployed on self-hosted Kubernetes and secured end-to-end.
My recent work focuses on building the platform layer that lets data and ML teams move fast without breaking things: reproducible deployments, automated security scanning, experiment tracking, and self-service tooling for engineers working on manufacturing-adjacent data problems.
My proficiency in each skill (drag to scroll)
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Used in most of my work
Education & Professional Certifications
University of Michigan Ann Arbor, 2011
GPA: 3.0/4.0
California State University Fullerton, 2024
GPA 3.9/4.0
In Progress
2026
Data engineering platforms and projects I'm working on
A comprehensive data engineering stack with modern open-source technologies
End-to-end reference architecture for manufacturing-adjacent data pipelines - events flow from Kafka through Flink into PostgreSQL, transformed by dbt, and surfaced in Metabase dashboards. Demonstrates how industrial sensor and operational data can be made queryable in near-real-time without a managed cloud data warehouse.
A Next.js internal developer portal giving engineers a single pane of glass over the entire data platform. Integrates Keycloak SSO, embeds live Grafana dashboards, and surfaces direct links to JupyterHub, Airflow, MLflow, and Metabase - eliminating bookmark sprawl and reducing time-to-tool for new team members.
Self-hosted MLflow deployment on Kubernetes backed by PostgreSQL and MinIO object storage. Provides a Databricks-compatible experiment tracking and model registry layer for JupyterHub workloads - enabling reproducible ML experiments and a clear path from notebook prototype to registered model artifact, without a managed cloud dependency.
Enterprise platform engineering projects and architectural decisions
Built a production-grade, self-hosted data platform on Kubernetes to eliminate dependence on managed cloud services for data-intensive workloads - delivering Databricks-class capabilities (experiment tracking, distributed compute, orchestrated pipelines) at a fraction of the cost, with full control over data residency and security posture.
Replaced a fragile, manually-managed Docker Compose environment with a fully declarative, GitOps-driven platform that any engineer can onboard to in minutes. The result is a reproducible, auditable data infrastructure that supports manufacturing-adjacent analytics workloads - with enterprise security controls typically only found in cloud-managed offerings.
Web properties and platforms I have built and operate
Internal developer portal for the Morning Star Engineering data platform. Single sign-on via Keycloak, embedded Grafana dashboards, and direct links to every platform tool - giving engineers one place to start their day.
Morning Star Engineering company website.
Morning Star Engineering ERP portal for internal operations management.
My reading journey through technical and professional development
This section will showcase the books I've read, organized by year, covering technical topics, leadership, and personal development.
I aim to read at least 4 technical and professional development books per year, focusing on software engineering, data and machine learning engineering, leadership, and personal growth.
© Copyright Tim Pomaville 2026
