Projects & Achievements
Here are some key initiatives I’ve worked on throughout my career in DevOps, platform engineering, and cloud infrastructure.
AWS Organization Tooling at T-Systems
Technologies: Python, AWS Lambda, DynamoDB, AWS Organizations, Billing Conductor, ECS, Security Hub, Jira API, Terraform
Custom automation tooling for a 350+ account AWS organization for the Generalitat de Catalunya, filling gaps where AWS doesn’t provide centralized management at org level. Notable tools built:
- Billing Conductor automation — cross-org Lambda that keeps every account assigned to the correct billing group based on tags, auto-fixes safe misplacements, and notifies finance and account owners of discrepancies
- ECS Instance Scheduler — DynamoDB-backed scheduling system similar to AWS’s official EC2/RDS scheduler (which doesn’t cover ECS) — scoped to non-productive environments via org unit IDs, services opt in via tags, scale to 0 overnight and on weekends, with replica counts stored to restore exact state the next working day
- Security Hub → Jira — monthly scan of critical and high findings across the org, creating Jira tickets assigned to the security owner of each department from a DynamoDB lookup table
Impact:
- Finance gets accurate per-department billing without manual reconciliation
- Non-productive ECS workloads shut down outside business hours, reducing costs
- Security findings automatically routed to the right owner every month
Kubernetes-Based Platform Operations at BlueLabs
Technologies: GCP, GKE, Argo CD, Terraform, Python
GKE environment on GCP with purpose-isolated clusters and GitOps workflows with Argo CD. Built a Python service to continuously monitor GCP spot node interruption notices and gracefully drain pods before node removal, protecting async workloads from in-transit data loss. Impact:
- Production workloads ran entirely on spot nodes without customer-facing reliability issues
- Graceful pod eviction eliminated the risk of data loss from abrupt node termination in async services
- GitOps workflows brought consistency and auditability to deployments across all environments
Observability Stack on Kubernetes at BlueLabs
Technologies: Loki, Grafana, Mimir, Prometheus, Thanos, Fluent-bit, Kubernetes
A collaborative team effort to build a centralized observability cluster that aggregated logs and metrics from the other GKE clusters. Each engineer owned a component — log collection with Fluent-bit shipping to Loki, metrics with Prometheus and Thanos shipping to Mimir, and Grafana for visualization. The whole stack deployed on spot nodes to keep costs low.
Impact:
- Full control over retention, alerting rules, and configuration — not possible with managed solutions
- Each team could instrument and visualize their own services without vendor constraints
- Faster debugging and incident response across all environments
Automated Domain Management at Scale at BlueLabs
Technologies: Cloudflare, BYOIP, Cloudflare for SaaS, Python, Retool
The engineering manager started the system; I took it over, developed it fully, and brought it to production. The Python service was designed from the start as an API — handling zone provisioning across Cloudflare, real-time availability monitoring, and automatic failover orchestration. The Retool back-office was one of its consumers, giving the team a visual interface to manage zone lifecycle. Also led the BYOIP initiative to bring a /24 IP block into Cloudflare.
Impact:
- Automated failover eliminated manual intervention during domain incidents
- Zone lifecycle tooling kept a 100+ zone estate manageable at scale
Self-Service Infrastructure Platform at Clarivate
Technologies: AWS, Terraform, Jenkins, Groovy, ECS Fargate, Backstage
A Jenkins-based automation platform with reusable Groovy libraries to deploy Terraform-managed infrastructure and ECS-based applications. Responsible for design, development, and rollout across teams via Backstage.
Impact:
- Enabled developer self-service for infrastructure and application deployment
- Reduced manual provisioning time by 80%
- Promoted infrastructure-as-code practices across teams
Cloud Cost Optimization at StubHub
Technologies: AWS, GCP, FinOps, CloudFormation, Terraform
Responsible for a FinOps initiative that cut annual cloud costs by 50%, saving over $1M — an achievement that directly contributed to a promotion to engineering lead. Full scope: defining a fully automated cloud infrastructure strategy and migrating legacy systems from on-prem to AWS.
Impact:
- Reduced cloud costs by 50%
- Improved reliability with 99.9% uptime
- Contributed to PCI-compliant deployments and international platform enablement
Technical Enablement & Team Leadership
Throughout my career I’ve often been the person colleagues turn to when they’re blocked — regardless of title. At StubHub I held this formally as a manager and mentor; elsewhere it’s been organic, helping senior and junior engineers alike work through problems, navigate architecture decisions, and raise the quality bar without anyone asking me to.
Want to Learn More?
Check out my GitHub or LinkedIn for more about my work and ideas. I’m also in Mastodon Looking to collaborate or need help with infrastructure challenges? Get in touch!