Experience

Replaced ticket queues with on-demand role vending. On GCP, Workload Identity Federation for keyless CI/CD authentication; on AWS, OIDC federation issuing temporary credentials constrained by permission boundaries.

A serverless, self-service system letting engineers patch EC2 instances with automated status reporting. Built on AWS Systems Manager, promoting validated patches from dev to prod with concurrency control. Patches 1,000+ instances monthly, cut the cycle from a week to one day, used by 10+ teams across 100 AWS accounts. The GCP version is in progress on VM Manager and Cloud Functions.

A Python and Jinja engine that renders complete GitLab CI and GitHub Actions pipelines from declarative YAML models.

A dynamic build fleet on EKS and GKE processing 1,000+ jobs weekly. Provisioned with Terragrunt, isolating 10 tenants via network policies and resource quotas. GKE autoscaling and Karpenter on EKS, with Buildah for rootless container builds.

Co-architected a hybrid image pipeline delivering 5 OS families and 20+ compute variants across AWS and GCP, including GPU images validated by automated test-pod deployment on GPU-enabled clusters.

Directed the automated release pipeline for monthly OS updates under QMS workflows with SBOM and CIS enforcement. Drove $10K in quarterly storage savings through automated image lifecycle and retirement analytics.

Automated rollout of AWS SCPs and RCPs and GCP Organization Policies through a Terraform, Python and Jinja pipeline. Removed public internet build dependencies by introducing Artifactory as a package proxy and compliance-artifact store.

An Amazon Bedrock-backed log analyser with a Python pipeline that cleans and deduplicates input to control token spend, and an autonomous code-review agent that flags anomalies in merge requests for human reviewers.

Lead an eight-person team, owning backlog, release management and stakeholder communication. Hired six engineers against an in-house technical bar.

Directed the GitHub Actions versus ArgoCD evaluation that selected the team's CI orchestrator.

Established documentation standards — product guides, FAQs and technical wikis — and delivered architecture briefings to 200+ cross-functional stakeholders.

Leading the GitLab to GitHub migration and pipeline re-architecture with a three-person team. Represented the platform at international PI planning in Roche Switzerland (2025).

Resolved execution limits on a TB-scale ingestion pipeline by re-architecting it as distributed serverless workloads, orchestrating independent read, write and process stages with SQS and Step Functions plus custom file indexing.

Internal data-aggregation tools on a fully serverless AWS stack, defined with AWS CDK and using Boto3 for cross-account API calls.

Primary on-call lead for high-severity production incidents. Drove root cause analyses, implemented preventative mitigations, and authored the SOPs and runbooks the rotation still uses.

A Jenkins job triggered by SQS on instance-termination events restores instances and databases with no manual step.

Shared libraries and multibranch pipelines for microservices and micro frontends, plus Kubernetes deployment and pod troubleshooting support.

Provisioned AWS with Terraform and CloudFormation using remote state and reusable modules, delivering a hybrid cloud migration with Packer-built AMIs.