We’re looking for a creative, analytical, geeky Linux system administrator to support advanced testing and performance systems of the Guidewire Software development team. We want an independent thinker who can make our internal servers awesomely high performing. We want an expert admin who can dive hands-on into system details but keeps the big picture in mind. This is NOT a traditional IT job. You won’t slog through user helpdesk tickets. You won’t focus all attention on uptime of old systems. Instead, top priority is maximum performance for our unit testing and performance testing infrastructure.
Do you love challenge? Do you like to dig deep? Do you have an overall geeky love of systems, tinkering, and browsing Linux log files? Can you tell the real reason for a Linux I/O wait? Do you know when to be safe with RAID 5 and when to use RAID 0 for speed? Do you love weird subtle mysteries? Can you make solid data-driven analysis? Can you look at output of the tools ifstat, iostat, and top, and know what would improve results on current hardware?
and analyzing perf data, to recommending major changes to processes.
- Make complex systems for QA and performance testing reliably superfast
- Constantly reassess and improve the developer infrastructure
- Work with development group to support their current & future needs
- Improve the automated imaging infrastructure
- Build out new servers and possibly entirely new sites in future
- Some hardware deployment and maintenance (racking servers, RAM upgrades)
- Evaluate processes and products, making well-documented recommendations
- On-call rotation with the team, 1 week every 8 weeks, it’s usually uneventful
- Occasional customer escalations relating to performance (about 3 per year)
- Diverse levels of work, from personally rewiring ports, to generating
Candidates must have ALL of the following technical skills:
Strong candidates have SOME of the following experience:
- Deep mastery of Linux, particularly with respect to performance
- Experience and extreme comfort with RedHat/CentOS Linux
- Comfort with at least one non-RPM-based Linux distribution
- Strong comfort with open source software
- Deep understanding of imaging & automated deployment for Linux systems
- Deep understanding of shell and/or perl scripting
- Deep understanding of common issues with HTTP, NFS, DNS, and DHCP
- Prior responsibility for locally-hosted mission-critical services
Strong candidates have ALL of the following personal traits:
- Virtualization, ideally Xen or other non-VMWare options
- Building custom kernel configurations for Linux
- Non-virtualized server management other than vCenter
- Software development agile methodologies such as skrum and kanban
- Some physical hardware deployment and maintenance
- Source control systems, at least as a user, ideally with “git”
- Bug tracking systems, ideally with “jira”
- Adminstering shared storage systems such as Netapp, EMC, SAN
- Hardware sizing and deployment for servers and direct-attached storage
- AIX on IBM Power hardware
- Multi-site applications, including disaster recovery and load-balancing
- Imaging and automated deployment of Windows servers
- Database and web-server scripting, such as PHP, modperl
- Hands-on admin of open-source block-level storage, such as iSCSI or FCoE
- Blade server management
- Previous work on huge projects
not important, but you must document what you did and why you did it.
- Comfort with bleeding-edge tools and processes
- Can flexibly prioritize immediate tasks and long term tasks
- Unafraid of big complex projects
- Service-oriented attitude (“How I can I help this team?”)
- Creative out-of-the-box thinking, not always deferring to industry approaches
- Comfort with and appreciation for software development teams
- Comfort with deep analysis and strategy with both big picture and critical details
- Comfort creating something truly new (not just maintaining old systems/processes)
- Can stay on task without getting stuck on irrelevant details
- Love of pushing the limits on what’s possible, pushing the envelope
- Works well with other people
- Can accept decisions about tools/processes that aren’t your preferred approach
- Comfort with data gathering and rational argument to make your case
- Can be clear and concise, especially when working with other teams
- Excellent verbal skills
- Excellent written skills, mostly for email and chat. Presentation skills are
Guidewire makes mission critical systems for multi-billion dollar insurance companies. To ensure our products are reliable, we strongly value test-driven development. We constantly brainstorm how to improve our coding and testing processes. Our software is great, the coffee is okay, and the buildings are decent enough. What’s special, though, is that we’re a team passionately focused on customer success and our long term goals. Talk to people here and you’ll immediately sense their enthusiasm. We’re proud of what we build and how we work together.
At the heart of Guidewire culture are three values:
Collegiality. We work together as professional equals, with a minimum of hierarchy. We strive for good processes and realistic assumptions to keep workload sustainable. We trust people to manage their own time. We seek input without regard for title. We encourage contrary opinions. We value work/life balance.
Integrity. We insist on truthful relationships with customers, prospective customers, partners, investors, and each other in everything we do.
Rationality. We communicate through clear arguments. We decide carefully based on factual evidence. We like calm, systematic, thoughtful solutions. We constantly improve processes.
A BA/BS in Computer Science/Engineering/MIS, or equivalent real world experience
10+ yrs. progressive experience in related positions.
Guidewire Software, Inc. - 20 months ago