As engineering teams scale — whether to keep up with technology demands, find better solutions or introduce new products — many are also diving in headfirst to modernize their approach.
When Chad Hoyt joined Braviant as director of data engineering, he was motivated by the fact that the financial services organization “truly values data as an asset.” Combining that powerful resource with other technologies such as machine learning not only makes credit lending more accessible, but also puts the company in a prime position to use more innovative practices to strengthen its database development.
One such practice for Hoyt’s team is the implementation of data lakehouses, which combine features of both data lakes and warehouses. This data storage expansion presented him with the challenge of restructuring the engineering team to rapidly scale its attendant growth. “To make this vision a reality would require additional support to maintain and grow the current platform while freeing up key individuals to work on the greenfield development,” he explained.
In addition to ensuring new hires had the opportunity to shadow well-versed developers and were provided with enough background information on the business and systems, Hoyt established a crowdsource-based data innovation workshop to “bring some fun collaboration to shared learning.”
Built In Chicago sat down with Hoyt — along with leaders from TransUnion and Clearcover — to learn how they’ve approached the difficulties of rapid scaling and what projects they’re ready to take on with their further-developed teams. Whether you’re an engineer wondering how you could contribute to a new team or a manager hoping to level up your leadership approach, the insights shared by these local professionals might point you in the right direction.
Why did your engineering team need to be expanded so quickly?
TransUnion’s developer experience team exists to enhance our engineering effectiveness throughout the company. In this fast-paced economy, innovation is key to success and increasingly in demand, which is why TransUnion needed to quickly engage top developers with a streamlined experience. As head of software engineering experience, my role focuses on building a strong team aligned around one shared vision. Our developer experience team encompasses all aspects of TransUnion, starting with onboarding, tools, systems and developer satisfaction. In simple terms, our customers and partners are our software engineers.
From building world-class tools enabling a streamlined developer experience to enhancing operational visibility and suggesting best practices for TransUnion’s engineers around the globe, my team wants the right thing to be easy to do. We are in a hypergrowth environment, and metrics showed that we weren’t increasing productivity as we envisioned. As we started gaining feedback through developer experience surveys and various micro-feedback loops, we saw the need to expand our team to make everything fast and friction-free.
We enable high-velocity development by adopting better solutions and reducing the overall cognitive load on the teams we support.”
What does the scaling process look like for your team? How do you adapt it to fit the company’s needs?
My team’s fundamental idea is to enable high-velocity development by adopting better solutions and reducing the overall cognitive load on the teams we support. TransUnion has a diverse product and technology stack with different types of build solutions, and through mergers and acquisitions, we’ve inherited various tools and workflows. Being able to scale horizontally is important as we build self-service tools and workflows to scale from 10 to hundreds of teams. Based on our key performance indicators and measured outcomes, we adjust and calibrate scaling needs to meet demands.
We also communicate our initiatives, roadmap, feature work and releases across various channels — such as chat groups and agile project management tools — and reach out to our developer community via engagement programs. While our team felt we were over-communicating, we actually found there was duplication of work being done in individual teams due to competing priorities and deadlines. This opened up an opportunity for us to investigate innersourcing our roadmap, allowing our developers to contribute to it, get early feedback on prioritization and help make decisions based on what we are or are not building.
What’s the most exciting project your engineering team will be taking on in the next few months?
There are at least a few projects I’m super excited about, but the most exciting one is automating change management in the continuous integration and delivery pipeline with auditing guardrails in place. TransUnion operates in a regulated industry space, and every change that goes to production is required to meet all the internal and external auditing and compliance standards and guidelines.
To shorten governance feedback loops and build a lightweight model around deployments and releases to higher environments, we are building evidence as a service. This service will extract all the auditing and compliance data from various tools in the pipeline — including dynamic and static application security testing, code coverage, and more — and provide the basis for creating audit dashboards. Integrating evidence as a service in CI/CD pipelines will help push changes to production without manual intervention. This is critical to make deployments continuous in our higher environments, eliminate manual steps for engineers and reduce context switching.
Why did your engineering team need to be expanded so quickly?
Upon joining Braviant, I was pleasantly surprised to find an organization that truly values data as an asset and believes in the promise of managing it as a product that can unlock business value. I’m continually inspired by the people and gem of a culture we have in place and humbled to have the opportunity to play a role in leading our data practice.
I was incredibly blessed to inherit a team of highly talented database development engineers. Previously, their focus had been to leverage PostgreSQL as the primary data store for managing both transactional and analytical workloads. They were doing a great job, but there were signs of fragility. As I met stakeholders in the business during my first 30 days, I found some amazing opportunities and interesting challenges to address. There was a great foundation in place with tech savvy business users primed to take new steps toward implementation of a strategy that expands into a data lakehouse and more innovative tools. Making this vision a reality would require additional support to maintain and grow the current platform while freeing up key individuals to work on the greenfield development.
Despite the complexity of our business, depth of integration and risks we had to navigate, our team has really risen to the challenge.”
What was the scaling process like, and how long did it take? Were there any hiccups or moments of clarity along the way?
With great ambition to grow the business, there was a need for new hires within each discipline. We hired two new managers and a data architect and transferred one manager and two SQL developers from another department while shifting a critical lead to the new data lakehouse, or DLH, initiative. This was very aggressive in a three- to four-month period. To scale this quickly required sharing business and systems knowledge with the new hires — myself included. We started with individual sessions to learn from existing developers and had new hires shadow others. We now have four disciplines that encompass online transactional processing database development, enterprise data warehouse database development, DLH development and data product management with a vision to evolve the practice to a more fluid, one-team organization in the future.
As for hiccups, I don’t think I fully appreciated the complexity of our business, depth of integration and risks we’d have to navigate, but our team has really risen to the challenge!
What’s the most exciting project your engineering team will be taking on in the next few months?
The DLH continues to be a major focus, and future use cases will help solidify its place in our organization. We have intentions to expand the offering to enable the business to do more with our data assets and unlock the potential of various business units. Through this initiative, we are bringing in tools that will help improve quality, governance and observability of our data to keep the business informed and build a trustworthy platform. I’m very excited about the prospect of this team’s future success and the potential I’m starting to see shine in each team member.
Why did your engineering team need to be expanded so quickly?
The Clearcover engineering team builds and supports all of the software and systems to provide our customers with the best online auto insurance experience in the industry. I am the senior vice president of technology and acting chief technology officer. I have been at Clearcover for nearly two years now, and our growth has been exceptional. In 2021, we expanded our coverage in eight states and saw 324 percent year-over-year growth in total policies in force. Everything we do at Clearcover is powered by software, and this rapid growth required us to double our engineering team from 50 to more than 100 people.
As we grow, we have continued to nurture current teams while forming new teams and initiatives. We rely heavily on Scrum as our execution framework, which has allowed us to hire new talent and provide them with the support of a team to onboard. While we do have documented onboarding materials, we are able to effectively onboard new members at the team level. With the support of their team and a paired mentor who provides a clear point of contact and day-to-day coaching outside of their manager, our new engineers are able to contribute on day one.
With the support of their team and a paired mentor, our new engineers are able to contribute on day one.”
Were there any hiccups or moments of clarity during the scaling process?
One of the challenges of doubling the size of our team was driving down friction in the engineering process, so we formed a developer productivity team to solely focus on removing that friction. This team has worked to improve our continuous integration and delivery process and automation. Recently, we added support for ephemeral testing environments that have allowed our teams to quickly test their changes in isolation and minimize a possible breakage of the main branch. Additionally, we have had to invest more in testing — specifically test automation.
Another challenge was the formation of new teams. Our approach was to seed new teams with a mixture of experienced engineers and new hires, which has allowed us to keep existing teams intact while supporting new hires with guidance from experienced engineers. Finding the right blend of leadership skills and personalities to allow a team to grow successfully requires continual review. We survey our engineers each quarter to measure our progress and employee engagement. This has kept us focused on what our team members need to continue growing.
What’s the most engaging project your engineering team will be taking on in the next few months?
I am particularly excited about building out all aspects of self-service for our customers. Insurance is traditionally a high-touch experience, and Clearcover has approached our position as a digital product to further separate what we offer from other providers.
Our mobile app is our customers’ main touchpoint with Clearcover, and we have an adoption rate of more than 85 percent. With the app, customers can add and remove drivers and vehicles on a policy and update their payment information and schedule. We have additional capabilities coming to allow our customers to use the app for all aspects of managing their insurance policy.
Additional improvements are also coming to our claims submission and tracking process, which support customers who have gotten into accidents, to make what can be a stressful time much more transparent. This builds on major investments we made into our claims processing speed last year. With the help of machine learning, our current record for paying out a claim is seven minutes. Features like these will go a long way toward allowing customers to fully engage in our self-service tools.
