Employment History

Joined E-commons when the process of setting up the hardware for the new SNIC AI/ML compute resource Alvis was fast approaching. Contributed mainly by taking on all network-related work for Alvis and re-engineering parts of the production network to increase resiliency.

Responsible for tools to manage the “sunset” phase of SNIC (National) Compute projects, to ensure that they get removed from centre storage according to national SNIC guidelines, with notifications to appropriate stakeholders and archivists.

Engaged with the IT-department developers to ensure our user integration can function more efficiently and reduce the workload for the team by automating tasks that were done manually.

Took on the role as implementation manager for the new ticketing system TopDesk that Chalmers provides together with the TopDesk team.

Maintained and planned upgrades/improvements to all parts of the networking infrastructure at the university, planning larger upgrades, ensuring that all staff and students experience uninterrupted service. I was responsible for transitioning the WiFi infrastructure from publicly routed to NAT, using a custom install of Pfsense in an active/active configuration. This involved transitioning all legacy tools, including the captive portal, over to pfsense. During the start-up phase I discovered a few defects in Pfsense that would make it less ideal for Chalmers’ needs, and had several patches accepted upstream. Most of these patches are also accepted in Opnsense, with additional patches that were refused by the Pfsense project.

Towards the end of my employment I was using network automation tools day-to-day as much as possible (Netmiko, Ansible, Napalm) as well as developing tools to help internal staff, general staff or even students to more easily use the systems we provide. A concrete example was a portal for staff to sponsor guests use of the network, by automating the creation of time-limited accounts. As part of this project I recommended the use of Git and Docker to the rest of the department, in order to manage updates and easier maintenance. Part of this was rolling out a Gitlab server to the department and integrating with our SSO (ADFS/SAML). On the tail end of this Gitlab together with Gitlab CI/CD was introduced as a tool for professors and researchers for building and automating their deployments, as well as being used in courses to teach CI/CD.

Lead the internal project tasked with replacing core networking infrastructure and transition fully to 100G.

At Double Negative I was the teams main Perl developer, responsible for maintaining most of the legacy Perl code as well as updating older tools to interact with modern monitoring systems. I rewrote the license monitoring stack from the ground up, using a modular system to collect current usage, and allow others to query this data using modern REST+JSON API methods. The data was also fed directly into Graphite.

During my time here, the team transitioned from an unstructured way of working into using agile methodology in most things we did, this ensured that we learned how to adapt quickly, but the most important thing I took away from this transition was learning how to appreciate the backlog and not let it be a stress factor. We built more reliable systems and at the same time reduced stress overall. As a team we managed to move away from legacy configuration management tools (Rdist for the curious) into more modern tools (Salt), using code review and Jenkins to manage changes in the environment.

Got their current OS-platform updated to a recent release as well as automated deployment and introduced configuration management systems.

I also developed disaster recovery plans, mapped out the current infrastructure, how it connected together and the critical points that were needed to ensure that the facility could recover in case of hardware failure.

Updated their virtual environment to a clustered KVM setup as well as virtualised all remaining services required to keep the facility running (authentication services, web frontend/databases and Gridengine for the cluster).

Optimised and installed storage hardware/software. Main production storage was four node GlusterFS. Secondary storage a three node Isilon OneFS setup.