Starting her career at Google and as an early product manager at Splunk, Divanny witnessed firsthand the rise of big data and automation. While automation is scary to some people because it’s being used to take the human out of the equation, automation’s potential as a democratizing force excites Divanny. She believes in taking a “human-in-the-loop approach” to automation that is interactive and operationalizes skills such as communication to help engineers solve their most complex problems.
M.R. Rangaswami: How is Transposit’s diverse team driving its success in building products and as a culture?
Divanny Lamas: It’s important for companies to keep challenging what makes a good engineering team and tech culture. We believe that a customer centric approach – one that emphasizes breaking down silos and tightening collaboration between different teams – is a powerful way to foster strong teams and fuel the innovation that comes from multiple perspectives.
What we continuously hear from the hiring process is that people want to work at diverse companies. They are tired of the status quo and outdated processes. The talent Transposit has been able to attract works differently from the aggressive engineering cultures of the early 2000s; for one thing, it’s a lot more collaborative.
We have a nearly 50% female engineering team and are led by an all-women C-suite. To us, diversity is more than looking at gender and race. We’re also evaluating candidates’ unique experiences and career accomplishments, and hiring for “culture add” versus “culture fit.” We’re big believers that everyone we bring on board has something exceptional to contribute, along with untapped potential. By focusing on mentorship and promoting from within, we help people’s distinct perspectives come alive in their work, and the whole team benefits.
M.R. What industry megatrends are impacting innovation in DevOps and holding back engineering organizations today?
Divanny: Businesses are currently under incredible pressure to remain competitive, or even stay afloat, while continuing to innovate. The pandemic has had a huge acceleration effect on digital transformation, with many companies moving their initiatives forward faster in six months than they had planned to do in six years. The explosion of SaaS and increased adoption of API workflows, microservices, and multi-cloud approaches, have all complicated DevOps, SRE, and other on-call engineering tasks. So at the very moment that software development needs to be moving faster, it’s being held back by ad hoc DevOps processes and outdated tools that were built before complex, modern stacks – and before DevOps philosophy – even existed.
Agile development promised faster innovation. But there’s a catch. Specialized QA teams are gone and engineers with varying levels of experience are now expected to support production systems. It’s incredibly stressful. Anyone working in operations knows the challenges of navigating and communicating information when the resolution clock is ticking. With institutional knowledge spread out across team members, during an incident there are too many people to message and it’s hard to keep track of communications. Slack data is unstructured, which makes searching and making sense of what happened before, during and after an incident painful and slow.
While incidents are an unavoidable part of modern operations, their accompanying chaos doesn’t have to be. We took a radically different approach to help DevOps teams – and really anyone involved in maintaining modern applications – regain the time and sanity they need to innovate. By automatically capturing and structuring human-generated data related to operations in a way that makes it accessible and contextual, Transposit makes it easy for engineers to keep track of the entire lifecycle of an incident, follow known procedures, collaborate and take action in a centralized place. We track this data to learn what works, building these insights into post-mortems and beyond. This speeds up issue resolution, letting businesses build new and better digital experiences for their customers with minimal downtime or degradation.
M.R. How do you see automation reducing the complexity of managing modern stacks and improving quality of life for engineering teams?
Divanny: The conversation around DevOps automation has been focused on end-to-end automation for a long time. The problem is that blindly “automating everything” without accommodating the role of the human operator leads to fragile systems and missed opportunities. Sometimes it stalls automation projects altogether, leaving companies with a handful of haphazard scripts that do more harm than good.
Done right, automation can be a democratizing force that levels the playing field. It makes it possible for new and veteran engineers to do their jobs more consistently, and with a lot less stress. For example, our platform uses human-in-the-loop automation to simplify and accelerate every step of the incident response lifecycle, from pulling relevant data from a monitoring tool, to posting stakeholder updates, to auto-remediation actions like reverting code commits or restarting a service. We let the human stay in command, while automating the toil of common steps that take time for humans to do manually. This combination of human judgment and workflow automation gives teams control without the hassle of the manual steps that most engineers are still taking today to resolve incidents.
But there is also another role automation can play in making DevOps easier and more sustainable. Why, in 2020, are we still relying on manual updates to ticketing systems like Jira or ServiceNow to keep track of what engineers are doing as they troubleshoot? Transposit’s platform automatically collects data across human and machine activities during daily operations and incidents, making it useful and searchable not just after an incident is over, but during the heat of the crisis. This speeds up time to resolution as engineers aren’t forced to dig through half-written documentation in a Confluence graveyard, and it also makes the post-incident reconstruction much more accurate than the current manual processes that most companies use. Having a complete, automatically generated picture of what humans are doing to interact with their systems empowers teams to prioritize development projects with a long-term strategic view that fully realizes the value of what automation can do for modern engineering orgs.