Mark Alberton Email & Phone Number

El Dorado Hills, California, Us

- Set up highly available, fault-tolerant infrastructure-as-code in AWS using Terraform, Packer, Ansible, and GitHub Actions.- Created a base layer of infrastructure for additional services to be built upon; including VPC, subnets, route tables, internet gateway, NAT gateway, and static IPs.- GitHub Actions deploys new code when committed to the correct branch.- Created an NGINX proxy. The Ansible script installed Certbot, generated certificates, and created the NGINX configuration files.- Created a VPN using OpenVPN and RADIUS authentication

Vancouver, Bc, Ca

- Installed a profiling tool (PyFlame) via Chef that took regular snapshots of the Python call stack, rendered it as a flame graph, and stored it in S3. Graphs were analyzed to find out potential areas of improvement in the codebase.- Added a lifecycle rule to daily EBS snapshots in order to decrease costs associated with storing them. - Used Packer to shorten build times by moving code out of Chef cookbooks and into pre-baked AMIs.- Built a Redis cache via Terraform and Chef to replace databases built ad-hoc. It contained a master node, disk backup node, and a user-defined number of read-replicas. Terraform built all the infrastructure present in AWS and ran the Chef bootstrap command to configure the Redis instances.- Migrated the entirety of the self-managed, ad-hoc TinyDNS system into AWS Route53 using Terraform.- Rebuilt the ad-hoc FluentD/Elasticsearch/Kibana log aggregation stack using Terraform and Chef. Auto-scaling groups were used to ensure the system maintained uptime.- Monitored system health using Datadog, InfluxDB, Grafana, and Kibana.- Took part in on-call rotation.

London, Gb

Responsible for upgrading both internal and client facing projects to make better use of cloud services. This involves writing infrastructure as code, automating processes and services, and architecting more cost-effective infrastructure. - Upgraded the internal Splunk cluster from an ad-hoc architecture to an automated, self-healing, multi-site cluster using Terraform.- Moved indexers into separate availability zones to ensure high availability- Migrated all indexers to auto scaling groups so they were self-healing and fault tolerant- Resized EC2 instances for the indexers and master node to be more cost-effective- Bootstrapped the Splunk AMI to upgrade to the version we use internally and then used Packer to build that into a new Splunk AMI with the correct version- Configured Splunk to archive frozen data and send to S3 for storage

Assisted in building and maintaining both traditional and cloud infrastructure/services to support the development of Fathom’s location services technology. The bulk of cloud services were provided by AWS and the majority of instances were using Ubuntu.- Provisioned resources inside of a VPC using CloudFormation (EC2 instances, auto-scaling groups, custom userdata scripts, ALBs, ElastiCache clusters, S3 buckets, IAM roles)- Established mappings to dynamically generate resources based on the environment- Deployed auto-scaling groups with launch configurations and scaling policies- Allowed rolling updates to ensure changes can be made as needed without disrupting the system- Secured instances behind ALBs by removing public internet access and requiring bastion hosts to log into specific instances- Secured bastion hosts with security groups that only allowed users from head office or over VPN to obtain access- Set up persistent VPN tunnels to VPCs- Created IAM users with policies to give the minimum amount of access to perform their tasks- Configured servers using Ansible to retrieve playbooks from a Git repository and run them locally- Stored private SSH keys for repository access in S3 buckets with custom policies that only gave access to specific instances- Benchmarked Redis clusters and EC2 instance types to find optimal balance of performance and cost- Attached VPN tunnels to the UniFi Security Gateway from the AWS Virtual Private Gateway- Developed Lambda functions using Python to run scheduled health checks on various services and another to remove untagged EC2 instances in the development environment- Wrote documentation on how to run OpenVPN using Docker- Worked extensively with Vagrant and VirtualBox using an Ubuntu image - Installed Ubiquiti switches (POE and non-POE) and security gateway- Ran/terminated Ethernet cable through office and tested throughput using iPerf3- Configured VOIP for office using Twilio, Asterisk PBX, and Fail2Ban

Built and updated the static site and Populace web application for Tantalus Labs hosted on Digital Ocean and Heroku, respectively (SSL-secured with CloudFlare). Populace is the web application used by Tantalus Labs for marketing and branding purposes. This is a Node JS application using the Ghost blogging framework.- Provisioned new Digital Ocean droplet to host the site. The site was secured with SSL using Cloudflare.- Used NGINX to add URL rewriting and a reverse proxy for the Populace app to route requests securely to the Heroku URL.- Migrated Populace to newly provisioned databases to conform to Heroku’s updated database structure.- Upgraded both the Heroku stack and Ghost framework as they were severely deprecated.- Responsive design that works on all devices and browsers.- Pricing data collected from competitors via scraping and statistically analyzed to determine fluctuations in market pricing.- Integrated Google Analytics for web traffic analysis.Technologies used: NGINX, Ubuntu, Digital Ocean, Heroku, Cloudflare, AWS S3, Node.js, Ghost, JavaScript, jQuery, D3, HTML5, CSS3, Foundation CSS, and Flexbox.

Web App Development

Master’S Degree, Epidemiology