A day in the life of a data scientist on the BrazeAIᵀᴹ forward-deployed engineering team

BrazeAIᵀᴹ forward-deployed engineering team data scientists are responsible for connecting organizations’ data with our system so that our AI agents can take the data, learn from it, and send the right information back to these businesses.

The forward-deployed engineering team is part of what sets Braze’s AI decisioning platform, BrazeAI Decisioning Studio™, apart. We help Braze customers build custom AI decisioning systems that are designed to deliver measurable—and sustainable—results that align to key marketing outcomes.

We do this by helping customers set up and launch BrazeAI Decisioning Studio™ for their unique business use cases, quantify the impact of AI decisioning on bottom-line business metrics, and iterate and optimize their campaigns on an ongoing basis.

I’ve been part of the team as a field data scientist for a year, offering customers partnership, best practices, and hands-on expertise. Here’s a peek behind the scenes of my day to day.

Like a plumber, I’m responsible for making sure things are flowing properly—in this case, the right data

When brands optimize their marketing campaigns with AI decisioning, the forward-deployed engineering team is responsible for the plumbing. Once we’ve worked with customers to define success and target their success metrics, we get to work on the backend. We get the data that we need from our customers, make sure it meets our quality standards, and ensure there’s sufficient data to train our AI agents so we can personalize campaigns and optimize a brand’s KPIs.

Our job is to connect organizations’ data with our systems so that our personalized AI agents can take that data, learn from it, and then send the recommendations back to these businesses.

We start with an exploratory analysis of sample data that the customer provides to make sure we have what we need. Then we configure the connections between our product and the brand’s data sources.

Once everything’s configured, we run statistical analyses and quality checks to make sure that everything’s up and running smoothly and that we’ve got best practices in place.

From there, we’re ready to go live, and the BrazeAI Decisioning Studio™ can begin generating recommendations and collecting training data to feed back into our AI models. 

As this is happening, we’re running analyses to make sure everything is working as intended and to measure the impact and uplift we’re generating for our customers. We report back on these findings and collaborate with our customers to help them get even more value out of the platform moving forward.

As a member of the deployment team, I help get the team launched and start personalizing campaigns and optimizing for their desired business KPIs. Once they’re consistently driving value, success teams manage the AI decisioning system on an ongoing basis.

I also work like a detective and inventor and put my sleuthing and experimentation skills to use

Getting AI decisioning to choose the right combination of personalization dimensions—such as the most effective marketing channels, message variants, creative, offers, and incentives—for individuals at the 1:1 level can be pretty complicated without the right resources in place. That’s where I come in. I help customers implement and use AI decisioning most effectively and make sure the AI acts as intended and delivers on the goals that have been set.

If customers aren’t achieving their desired outcomes, I dig into the data and figure out what tweaks and improvements can be made. As I do my detective work, I also test out different hypotheses to determine the root cause of issues.

I work closely with other members of the forward-deployed engineering team

We have a tight knit unit that consists of a forward-deployed data scientist like myself, an engagement manager, a customer success team member, and a tech lead, who all work closely together to help our customers achieve success with our AI decisioning platform.

On a typical day, you’ll find me…

…in customer meetings providing status updates.

…reviewing internal dashboards, configurations, and alerts to make sure that everything is running smoothly.

…getting new customers set up, reviewing their data, running analyses, and getting them going with BrazeAI Decisioning Studio™.

…testing hypotheses to get to the root cause of challenges customers are facing.

…figuring out new ways for the team and our tools to work more efficiently and catch and resolve problems faster.

The best part of the job is providing value for our customers

Historically, it hasn’t been easy for marketers to tie their efforts to real business impact, but that’s changing with AI decisioning. In most cases, when we launch a new initiative for a customer using our platform, we’re able to see the uplift it delivers very quickly and that’s incredibly satisfying.

Because of the way our product works, we’re able to say, “Okay, we launched this new campaign and we can now accurately help to quantify how much each campaign contributes to revenue.” That level of direct correlation is honestly something I haven't seen elsewhere in marketing.

Why it helps to have a human like me in the loop

There’s no one-size-fits-all cookie cutter AI decisioning system that will work for all industries, brands, and marketing use cases. That’s why it’s helpful to have a human in the loop to connect the dots between what a company’s goals and objectives are and how the AI works. That’s when I can jump in and figure out how to configure our system in ways that are optimal for a company’s specific needs.

That’s a wrap on my day-to-day

Want to see what our team of forward-deployed engineers can do for your marketing efforts? Learn how we can help you get more value out of AI decisioning by handling the heavy lifting of implementation and see the steps we take to set your team up for success.