Fairwinds built the Insights SaaS platform to provide visibility into Kubernetes clusters and provides professional services to manage Kubernetes clusters for clients on various cloud platforms. After seven months at the company I was promoted to a custom-built role to lead an SRE team responsible for the technical implementation of Insights in clients’ Kubernetes infrastructure, and to also guide clients in optimizing the security, efficiency, reliability, and cost of their Kubernetes clusters. I also served as a voice of the customer to product teams. My work included advising clients on security vulnerabilities in CI processes, Kubernetes configuration,right sizing, navigating issues across vendors (i.e. AWS, GCP, OpenShift), etc.As a leader, I was responsible for ensuring the technical capabilities of our teams (SREs, sales) by facilitating biweekly learning sessions, 1:1 check ins, and ad-hoc enablement sessions on Kubernetes best practices and the Insights platform.Project Highlights: Developed Open Policy Agent policies for clients and for an internal library to set guard rails for Kubernetes with the Admission Controller. Created a project board and ticketing system in YouTrack to receive customer support issues via email to replace Slack-based support.Identified an SRE pain point related to crash loops and recommended build of a tool to detect behaviors and adjust memory. Provided specs to developers for the tool that was built in Golang and released in six months. Represented Fairwinds as a speaker or panelist at several conferences, events, and the Klustered livestream. Led an internal book club to discuss “Radical Candor” that focused on giving productive feedback, with participants ranging from customer success leadership to C-suite. MEDIA: https://youtu.be/_GZ8UuEZxbY?t=3301

I was initially hired to manage and monitor ~100 Kubernetes clusters (on AWS, GCP, Azure), manage infrastructure in Kops and Terraform, track down issues in Datadog, and provide on-call support for production issues to include auto scaling problems, networking debugging, cloud issues, OOMKills, performance bottlenecks, etc. I was also responsible for onboarding new engineers and providing internal learning sessions on various topics (CAP theorem, sys calls and eBPF, writing policies with the Open Policy Agent, etc.).Other work included maintaining the internal tooling in Python to manage clusters, create infrastructure, handle permissions, and perform a variety of other helper functions for managing infrastructure.

The Pew Research Center is a nonpartisan fact tank. I joined as Pew’s only DevOps resource to build and maintain infrastructure to support data scientists in data collection, with responsibility for scaling data collection on Kubernetes, implementing CI/CD workflows to expedite development and deployment, and automating tasks to reduce points of friction.Project Highlights:* Built and owned a Kubernetes cluster on EKS. Migrated applications, configured Helm charts, connected RDS/EFS, configured network policies, linked AWS IAM permissions to Kubernetes service accounts, set up routine etcd backups, and established a reverse proxy NGINX-Ingress controller for routing traffic to web accessible services. * Reduced deployment times and maintenance overhead by introducing CI/CD pipelines on ArgoCD/Github Actions.* Developed Python services and Lambda functions to read/write/pass data from DynamoDB to data collectors, clean EBS snapshots, build AMIs on a schedule, with Slack notifications as needed. * Reduced downtime, cost, and workload by setting up EFS, migrating data from a single NFS node, and reconnecting everything to EC2 instances and Kubernetes cluster.* Authored documentation for manual tasks and facilitated training on Kubernetes concepts. * Identified developer pain points and architected a scheduling system to standardize data collection across projects using DynamoDB, Lambda functions, Elasticache, and Celery in Python. Configured Cloudwatch alarms to send notifications and error messages to Slack.* Automated user creation in Ansible to trigger updates for users and their home folders on SFTP/NFS servers.* Introduced blue/green deployments with Terraform for critical infrastructure to reduce deployment downtime.