Streamline migrates AI-powered marketing platform for startups to Google Cloud from AWS

Startups and small-and-medium-sized businesses (SMBs) have traditionally faced challenges accessing and affording the technology used by large enterprises to scale and grow their companies in the cloud. 

Cameron Sechrist, CEO and Co-founder of Streamline, set out to develop a platform to fill this gap. “We saw a need in the market for an innovative marketing operations and business automation platform tailored to startups and SMBs that could provide the necessary AI-powered tools for these companies to market, sell, automate, and manage their businesses,” says Sechrist. “Our goal is to put the power of technology used in larger organizations into the hands of everyone and within their budget.”

After countless hours of planning, building, and optimizing 38 million lines of code, Streamline recently launched their marketing platform, empowering customers to create campaigns, ads, and emails with efficiency and ease. The solution also provides payment optimization, product management, workflow automation, and powerful reporting capabilities. 

Business challenge 

A key component of Streamline’s marketing software platform is the artificial intelligence algorithm for the virtual assistant, Tyra, which can help generate content, write campaigns, and provide customers with business insights. “When we were gearing up the platform, we were really struggling with the infrastructure,” says Sechrist. “Tyra has to have access to a lot of GPUs and other resources to run effectively.”

When trying to train Tyra models using the company’s original cloud provider, AWS, the developers were blocked as they tried to access resources. Workflows froze, and some products broke due to data overload. 

“At that point, we realized that we needed to find another provider so that we could sleep soundly. We were looking for a cloud provider that could help us if there is an outage not caused by Streamline,” says Sechrist. “I wanted to be able to talk to someone and get updates on critical issues. Putting in a support ticket and waiting 72 hours for a response was not an option.”

Solution

Streamline had already been using Gmail, part of the Google Workspace suite of tools, for their company’s email and collaboration needs, including Calendar, Meet, Chat, and Docs. Because of this existing business relationship, Sechrist reached out to Google Cloud to discuss the issue of GPU access.

A Google Cloud representative responded quickly. “The fact that I was actually able to talk with a human being was mind-blowing,” says Sechrist. Google Cloud introduced Streamline to the powerful capabilities of NVIDIA A100 and V100 GPUs on Compute, with configurable, high-performance virtual machines (VMs) to accelerate their most demanding workloads. 

Google Cloud also recommended that Streamline collaborate with SADA, a multiple-time Google Cloud Partner of the Year. SADA worked with Google Cloud to whitelist Streamline’s account so they could get access to a higher quota of GPUs to support the Tyra AI agent. SADA also assisted with their migration to Google Cloud, helped configure their infrastructure, and delivered other services to the organization.

“SADA helped me and the Google Cloud team figure out if Vertex AI would work with our platform,” says Sechrist. “It’s really helpful to get that level of insight, which I’m not going to get as a platform user. It’s only working closely with those teams who use the platform every day in multiple use cases that you get a true breadth of knowledge.”

I was sold on working with SADA just from how responsive and helpful everyone was in answering our questions and understanding our needs. They operate with the notion that if we succeed, they succeed. That was refreshing.

Cameron Sechrist | CEO and Co-founder of Streamline

Impact

As a result of collaborating with SADA, Streamline was able to finish their migration to Google Cloud from AWS and evaluate and develop an infrastructure roadmap going forward. “We saved at least 40 man-hours on our infrastructure strategy by collaborating with SADA,” says Sechrist. “In addition, our migration to Google Cloud from AWS was quick and easy, saving us 30% on infrastructure costs in the process.”

With more access to NVIDIA GPUs, Streamline has also enjoyed a quantum leap in AI training efficiency. “It would’ve taken around 16.5 years to train Tyra, our core AI algorithm, using AWS infrastructure,” says Sechrist. “We were able to cut that down to one month with Google Cloud.” Streamline now enjoys on-demand access to 100 NVIDIA A100 Tensor Core GPUs with memory bandwidth of over two terabytes per second to run their large models and datasets. 

SADA has been and continues to be our safety net. They are like the Sherpas of Google Cloud and provide top-notch product solution consulting, which allows our engineers to move beyond infrastructure concerns and spend more time delivering innovative features for our platform.

Cameron Sechrist | CEO and Co-founder of Streamline

Overall, SADA helped Streamline:

• Migrate their AI platform to Google Cloud from AWS 
• Gain access to 100 NVIDIA GPUs running on Compute
• Optimize Google Cloud resources to meet infrastructure demands
• Save 30% on infrastructure costs