As a senior manager in a SaaS environment, a trend that I see starting to dominate the DevOps hiring landscape is the widening gulf between technological innovation and those who have the exposure to it. For example, I recently saw a job listing on Craigslist, and one particular phrase caught my eye:
"The CTO appreciates a varied background with a CURRENT SKILL SET for the above skills and he has been frustrated with other vendors who send over candidates with older skills".
That posting alludes to the challenge that faces many DevOps job seekers. Staying current in arguably the most rapidly-changing technical discipline can be nearly impossible. To remain ‘cutting edge’ either requires a healthy IT budget, especially in the realm of hardware or cloud hosting exposure, or alternatively a technical environment that is small enough to allow for rapid prototyping and experimentation. And, if you’re a DevOps-focused engineer in the boom/bust cycle of SaaS startups who gets laid off or prove unable to adapt to more sophisticated environments the odds of getting your foot in the door at another company can go down significantly. For a technical professional, just six months out of work can have a devastating effect, introducing the risk of potential employers questioning their inability to find another job.
The DevOps recruitment paradox
This is not just a problem that "rank and file" engineers have to face. In fact it's a well-known phenomenon among technical managers that after a certain period you may lose touch with the requisite tech innovations, causing your to suffer. If laid off, there is the risk of having to start at the bottom as a junior engineer again in order to regain the relevant skills. Many hiring managers have gone down the route of rejecting "over-qualified" senior managers in favour of less-seasoned, more current (and often cheaper) candidates, and in the realm of DevOps, where technology in the deployment, configuration and virtualization space is changing remarkably quickly, the "out-of-date" card is even more likely to be played.
The paradox of this situation is that filling DevOps positions is already notoriously difficult. Even without bias against technical currency, there is a critical shortage in the market. And, many companies know they need DevOps in order to deliver products more quickly, but don’t know how to begin that journey. Yet, when they do open those positions they often throw candidates into the digital dust bin simply because their resumés are lacking the most current keyword like “Docker” or “Chef.”
Asking for specific DevOps qualifications is counter-productive
For companies wishing to start the DevOps journey, there’s the added complication of hiring when you’re not able to vet their technical qualifications. It’s wise to consult a more technical colleague when drafting job description requirements and conducting a background check on potential new hires in this field. For instance, asking for very specific qualifications and experience in applications or hardware may be unnecessarily restrictive, and someone with more expertise could probably identify equivalent experience. For example, someone who has very strong experience in Ubuntu Linux will probably do quite well in Red Hat Enterprise Linux (RHEL) after a brief acquaintance period. Incidentally, some of the best Linux admins I have ever managed with were experts in Sun Solaris, which has primarily been the domain of very large Unix-based shops. Connections like these would be difficult to make without a technical background. They also conceal candidates who might be excellent DevOps engineers but have never held that title or specific responsibility.
Interestingly, a trend I’ve spotted is the number of new-to-market DevOps engineers who have worked exclusively with and for Amazon Web Services or another cloud provider. So for those cost-conscious businesses seeking to co-locate their infrastructure in datacentres, be ready to cope with potentially lengthy transitions of employees in an unfamiliar work environment or else hire someone who has experience enduring the burdensome tasks of, say, cabling and organising racks of equipment. I include this as a case where skills that might seem very applicable most often are not.
Certifications are not a replacement for knowledge
Turning briefly to certifications (which are currently missing in pure DevOps) - although they have their place in recruitment, I often fail to see their value in predicting job performance unless you’re hiring for a very specific job function. They are more of a way to gain credibility without necessarily having the work experience to back up an application. Studying for and passing exams has its benefits, but it is definitely not a substitute for first-hand knowledge and experience of the career.
In my experience as a manager, the best tech staff I’ve had the privilege of working with started out as rogues or hackers with little interest in formalising their approach or codifying their skillsets. They learned by doing, and more importantly by doing it wrong. Hands-on, empirical learners tend to connect cause and effect in a very direct way, while those with more theoretical, less practical knowledge tend to refer to documentation or external support resources when trying to solve a problem. This is not necessarily a disadvantage, but definitely something to keep in mind when hiring. The hands-on folks are typically better builders, while theoretical engineers are better maintainers. This split is something I predict will be addressed more and more in future years.
Take a leap of faith for the right hire
As we see the job market and recruitment processes evolving, recruiters need to be adapting their strategies to keep up with today’s ever-changing DevOps landscape and demand for skills currency. This might mean taking a bit of a leap of faith to find the right hire, because there are many people just a few steps behind the curve who could quickly be brought up to speed when given access to cutting edge technologies. It’s worth adding that you needn’t say no to senior candidates, as they can be just as hungry for the opportunity to re-immerse themselves in current technology.
The main thing to remember is that DevOps as a discipline allows for a wide spectrum of cross-functional skills and experience and, in an increasingly diverse hiring pool, a ‘one size fits all’ approach is not appropriate. So before you hastily fill that position, make sure you really know about the job you are hiring for or consult those that do.
About the author
Jason DuMars is Director of Technical Operations at transactional and marketing email provider SendGrid. Jason is a passionate evangelist for the agile enterprise, DevOps and seamless agile engineering practices. Jason writes and speaks about agile engineering regularly, and is a member of the DevOps Enterprise Culture Working Group.