Gordon Murray

Took full ownership of AWS infrastructure across multiple AWS accounts and regions previously managed by contractors, establishing standardised Terraform/IaC, reusable modules, deployment patterns, and operational visibility. Rebuilt a manual disaster-recovery process as a complete Terraform-defined recovery region reducing recovery time from weeks to hours. Resolved DynamoDB hot-partition issues by redesigning partition keys and access patterns, eliminating throttling in production workloads.

Redesigned OpenSearch architecture from a shared multi-tenant index to tenant-isolated indices with improved shard configuration, reducing query contention and improving reliability for high-traffic workloads. Founded and now lead an internal AWS operations and observability function using CloudWatch, New Relic, and Terraform, improving alerting coverage, incident response, and system visibility across engineering teams. Coordinated infrastructure and platform work daily with US-based leadership and with developers in Germany and India, across US, European and Indian working days. Reduced monthly AWS spend by double-digit percentages through rightsizing, Savings Plans and reserved-instance planning, and ongoing cost-governance practices. Introduced tagging, guardrails, and automated policy checks to reduce configuration drift and support compliance across AWS accounts.

Reverse-engineered an undocumented Jenkins/Bash/S3 deployment system and rebuilt it as a GitLab CI/CD workflow using ECR, Terraform and ECS, with Octopus Deploy for release management, turning fragile manual releases into repeatable, version-controlled deployments. Partnered with backend engineers to containerize internal APIs and background workers, improving deployment cadence, service isolation, and production reliability. Supported colleagues in scaling NVIDIA Triton Inference Server and OpenCLIP models for production use.

Designed and deployed a production Change Data Capture pipeline using Apache Flink, MariaDB, Debezium, Kafka, and KSQLDB to capture project-management activity across 600,000 tenants in US and EU regions, optimising application caches and enriching the data lake for analytical and operational needs. Handled high-volume data streams with strategies to ensure data integrity and availability, including diagnosing replication lag and consumer backpressure under production load. Instituted data governance practices preventing transmission of personally identifiable information, architecting and enforcing data privacy measures for GDPR compliance. Worked as part of a dedicated senior platform team focused on improving integration and collaboration across departments.

Founded the SysOps function as the first dedicated infrastructure hire and grew it to a seven-person team of two DBAs and five SysOps/platform engineers. Supported growth from three customer-facing products in one AWS region to five Go-based multi-tenant SaaS products across two AWS regions. Owned production reliability, patching, infrastructure operations, monitoring, alerting, and a 24/7 on-call rotation for customer-facing systems. Worked with remote colleagues and teams in the US, Amsterdam and Australia across a distributed engineering organisation. Led migration from EC2/ECS workloads to Kubernetes, including Kops and later EKS, introducing Helm based deployments, autoscaling, and GitOps workflows.

Operated production database, cache, search, and traffic-management layers including MySQL, MariaDB, Redis/ElastiCache, OpenSearch/Elasticsearch, HAProxy, and relational replicas. Delivered repeatable database maintenance and zero-downtime schema-change workflows using ProxySQL, Percona tooling, and pt-archiver under sustained production pressure. Introduced HashiCorp Vault for centralised secret management and supported the operational controls needed for SOC 2 and ISO 27001 certification.

Adopted Cloud Custodian and policy automation to detect exposed assets, enforce tagging, reduce configuration drift, and improve cost reporting, alongside regular pentests and automated scanning. Built a custom domain service supporting customer custom domains with HTTPS.

Implemented centralised logging, alerting, runbooks, emergency procedures, access-control processes, and offboarding controls.

Operated Datadog and Sentry for application monitoring and error tracking across the customer-facing products, and led internal work to reduce their cost as usage and ingest volumes grew.