If you’re in the financial services or technology industries, you’ve likely heard of Vestmark. They provide software and services for some of the world’s largest asset management companies.
However, what you likely don’t know is that they’re also a visionary engineering team. We sat down with Freedom Dumlao, Vestmark CTO, to talk about how using Heroku was like unlocking the company’s latent superpowers, radically changing how they think about software development and setting them up to win in the age of AI.
A visionary engineering team unlocks superpowers with PaaS
Vestmark has no shortage of brilliant engineering talent, executing new ideas and cutting-edge methods despite a 20-year-old technology platform. However, this eagerness to develop new capabilities for clients led to organizational friction as different teams were all doing different things in different ways. One team was building things in Go, another in .NET, another in node.js, with each team choosing their own tech stack for different (and very good) reasons.
Therefore, Vestmark realized they needed to look ahead, think beyond what they were building, and consider what they would end up operating in the future. Building software is only the beginning of the story—after that, you have to operate, update, and maintain it, and for a much longer time than it took to originally build it. Each different tech stack was easy to choose in the short term but would increase the company’s long-term operational burden.
Client-centric solutions, not self-imposed problems
Vestmark needed to create a solution to bring all these disparate teams and their tech stacks together. Leadership was impressed by internal efforts to create custom web frameworks for this purpose, but it raised a worrying question: “Are we creating new problems for ourselves that require us to invest time and resources into solving them?” Ideally, Vestmark wanted to put 100% of its time, effort, and resources into solving problems for clients and staying competitive. They needed to find one place where everyone could work on the same code base and bring all their capabilities together, allowing engineers to focus on solving problems for clients instead of gluing everything together internally.
Complexity cannot be destroyed, only moved. The question becomes: ‘Where will it exist in your stack?’ What I like about Heroku is that while there’s a ton of complexity, I don’t have to think about it. We moved the complexity to where Heroku could deal with it so we could focus on solving problems for customers.
– Freedom Dumlao, Vestmark CTO
Overcoming organizational challenges
The first step to overcoming these challenges was organizational. Some people were working on their core technology, while everyone else was organized into smaller teams working on new projects, each in their own way with their own solutions. Vestmark grouped all the different, disconnected engineering teams into a single new entity called “the foundry” where the best product ideas could be forged into actual real-world products. This team came together very quickly with the mission of building, deploying, and validating new products.
This organizational change removed a lot of friction, but it also came with the mandate to go and solve problems. While the cloud engineering team had heard of continuous integration, it’s another thing to actually support building and deploying something 3-4 times in a single day. It was a new concept to execute, and getting to the point where they could rapidly deploy and iterate wasn’t feasible with their current setup. They needed Heroku.
Choosing Heroku was easy—using it was even easier
Choosing Heroku was easy, given that Vestmark was already a customer of AWS and Salesforce. With the vetting done by AWS and Salesforce, not much more was needed to get Heroku through their internal compliance mechanism.
Getting started was even easier: leadership logged in, entered credit card information, and got their code into the cloud that very same day. Configuring SSL was also easy, thanks to Heroku’s auto-configuration tool. With a single press of a button, they had a Postgres database they could connect to, and they didn’t have to think about deploying it. The process of building and deploying apps was frictionless – all done from the CLI or web console.
Heroku was an easy choice, and what made us stay with Heroku was how immediately successful we were with it.
Having a team of people who had never seen or used Heroku before, successfully managing Dynos and deploying software in a matter of days? There isn’t another cloud story that’s quite like that.
– Freedom Dumlao, Vestmark CTO
Deploying code in 43 days instead of 2 years
Vestmark set itself an ambitious goal: use the new tech stack to create a new product, from first line of code to deployed and in the hands of the customer, in weeks. Previously, this could have taken years. Despite being new to Heroku, they were able to deploy rapidly, build a customer feedback loop, and accomplish their goal in 43 days.
I told people I wanted a ‘Heroku-like deployment experience,’ meaning I could get into my CLI, enter a couple of commands, and my code is now running in the cloud and available to be accessed by my customers and product teams.
It turns out that if you want a Heroku-like deployment experience, the easiest way to get that is to use Heroku. So that’s what we did—we chose Heroku and found success with it almost immediately.
– Freedom Dumlao, Vestmark CTO
Unlocking superpowers with Heroku
Using Heroku created a new way of thinking at Vestmark, completely unlocking the imaginations of their engineers and product teams. Now, Vestmark can find value by spending a few weeks and a few thousand dollars instead of a few years and a few million dollars, completely changing the way they think about their product roadmap.
This wouldn’t be possible if they weren’t able to build and deploy quickly. Today, their teams can get a minimal version of working software in front of a client almost as quickly as they could build a Figma wireframe. Once clients could actually interact with software instead of concept art, Vestmark noticed a huge increase in client confidence.
I can build a product, get it in front of a client, and get the client to tell me if there’s value right away. If we’re on the wrong track we can pivot before we’ve spent two years developing a product. It’s like a superpower that we’ve unlocked.
– Freedom Dumlao, Vestmark CTO
Bringing the engineering and product support teams together
It’s important to Vestmark’s engineering team leadership that the engineering team is part of their software’s support group. When there are layers between the engineering team and managing the software in the cloud, it can be difficult for them to participate as part of the team that keeps everything running well. With Heroku, the engineering team has the exact same access as the support group, allowing them to see all of the weird and wonderful ways customers can break their software and be on-hand to help fix it.
When we’re able to respond to a bug or failure in a matter of minutes before anyone notices, my quality equation changes dramatically. I’m able to feel really confident putting something out there, knowing that if we publish a bug we can get it resolved before it impacts the client.
– Freedom Dumlao, Vestmark CTO
Vast savings for Vestmark
Utilizing Heroku, Vestmark understood they could focus on continuous integration and deployment without investing in a massive DevOps team. Instead, they utilize their current team to develop new products and services.
Easily, we’re somewhere between 60% and 80% time savings, using Heroku and Ruby on Rails together. We’re spending far less time figuring out how to integrate, deploy, test, and get feedback. All those questions have disappeared. Now it’s just write code, see the results, get feedback, and do it again.
– Freedom Dumlao, Vestmark CTO
What’s next for Vestmark?
Vestmark believes that agentic AI will be the toolset that defines their industry. They are excited about the prospect of using Heroku Managed Inference and Agents to simplify AI/ML model deployment with a production-ready API and Agentforce. They believe that the companies that will thrive in the age of AI are those that build what wasn’t possible before today’s powerful AI models.
Launch your next project with Heroku
How Vestmark did it
- They use a Heroku Postgres database that provides safe, stable, and frictionless connections
- The team just started using Agentforce, getting agents up and running with a simple CLI command
- Heroku Managed Inference and Agents help get models running and attached to their apps, saving time and headaches
- They securely add and manage credentials using Heroku’s configuration systems via Dynos