DevOps Developer in Sydney, Australia

Leading a team of three, I built all necessary code to allow the customer, a large insurance company, to manage their Akamai CDN via CDK for Terraform.

This involved writing CDKTF constructs to make managing Akamai via Terraform simpler and a full CI/CD pipeline for automated testing and deployment of changes through relevant environments (dev, test, production).
The CI/CD pipeline was built using GitHub actions.

In addition, I wrote code that utilized the Akamai API to convert existing 'clickops' created properties to the newly built infrastructure-as-code platform.

The security team at an international airline was concerned that no intrusion detection system (IDS) was in place for network traffic within or exiting AWS.
They considered extremely expensive IDS 'devices' for capturing all VPC traffic.

I wrote reports and presented them to senior stakeholders to help them understand their requirements could be met and even exceeded by using AWS GuardDuty, a solution available at a fraction of the cost of the IDS solution they were considering.

I oversaw implementing GuardDuty across the estate of several hundred AWS accounts, creating automation to ensure that newly created accounts were automatically enrolled in GuardDuty. The automation I created also ensured that SoC and those in charge of the infrastructure findings were related to were altered.

Automation was performed via a combination of AWS CloudFormation StackSets and AWS Lambda, with functions written in Go.

When I joined Qantas, they used a peering configuration to provide interconnectivity between their various VPCs. Also, they attached AWS Direct Connect to individual VPCs manually. This was error-prone because of the manual steps involved, and it restricted their AWS growth because of rigid limitations on AWS peering.

I designed and implemented a move to AWS Transit Gateway. This removed bottlenecks, relieving the immediate pain point. The implementation provided automation for newly created accounts, utilizing AWS Lambda and AWS CloudFormation to connect new accounts appropriately without requiring any human interaction.

Transit Gateway additionally provided a path for multi-cloud adoption, with Google Cloud and Azure peering providing significant extra flexibility for the organization.

Qantas was hit particularly hard by COVID-19 because Australia imposed a strict quota on the number of international arrivals for almost two years.

Working with another team member in charge of shutting down unused and right-sizing AWS instances, I focused on less obvious methods for finding cost savings.

I was able to find several million dollars annually in additional savings, over and above those from more traditional cost optimization techniques.
I did this by utilizing Python and Boto3 to interrogate several hundred AWS accounts and devising innovative strategies to save costs.

These included:

• Finding and removing unused RDS and RDS Aurora instances.
• Aligning AWS RDS and RDS Aurora instance type used across accounts and teams, enabling significantly higher reserved instance coverage.
• Finding and removing unused EBS volumes.
• Converting provisioned IOPS EBS volumes to gp3 volumes, which provided the same performance at a much lower cost.
• Converting gp2 volumes to gp3 volumes
• Finding and removing unused ELBs.

All of these strategies were implemented in a fully automated manner, with no more than 60 seconds of downtime for any individual change.

A UK cosmetics producer and retailer, LUSH, needed to migrate away from their AWS eCommerce environment, which had significant availability problems.

We were given an insanely short timescale of 30 days, so I led a team of ten who managed to refactor and migrate the environment in just 22 days. The efforts were worthwhile as LUSH had their largest ever daily sale just one month later, with the new infrastructure providing 100% uptime throughout this period.

A TV streaming service for ITV plc, the largest independent TV broadcaster in the UK. I architected and led the implementation of the AWS environment. This included a highly-available CI/CD solution that handles millions of concurrent viewers. The solution was featured in the opening keynote for the AWS Summit UK.

I led a team that successfully migrated Oddschecker from a traditional on-premise server environment to a fully containerized environment running on Google Kubernetes Engine (GKE) and using Google Cloud SQL.

As well as doing design and hands-on work, I also provided training to upskill the in-house team to support the environment.

These are the slides from a conference talk given in 2015 that demonstrate my lengthy experience with Terraform.

The talk introduced infrastructure as code more generally and talked about the specifics of managing AWS via Terraform.

I led a team migrating a major London fintech company from co-located data center hardware into AWS.

I refactored their environment into highly available autoscaling infrastructure and migrated from an active and passive MySQL configuration to Amazon Aurora.

The customer enjoyed significantly better uptime and 30% lower costs. Please see the AWS case study in the project URL for full details.