decorative pattern
decorative pattern Case Study is a fast-growing startup specializing in AI-powered image processing and optimization solutions.

Let's Talk a case story

The Client is a fast-growing startup specializing in AI-powered image processing and optimization solutions. The company leverages advanced computer vision and machine learning algorithms to develop solutions to automate various image processing tasks, such as object recognition, image classification, and content moderation.

AI Image Processing
Main Technologies
AWS, Kubernetes, Spot Instances, GPU
Date of Project

The Challenge

As the company grew and started onboarding its first enterprise clients, the team realized that the existing EC2-based infrastructure didn’t meet the demands of large enterprise clients who expect high availability and production-grade reliability from their 3rd party software providers.

To address this challenge, Bria considered moving to Kubernetes but didn’t have the in-house expertise to make the migration quick and with zero downtime. A friend introduced them to Opsfleet, a partner that could help jumpstart their move and provide ongoing support for architectural decisions and implementation.

“After signing a deal with one of our first big clients, we tried to scale our infrastructure with AWS node groups and additional ad-hoc fixes but quickly realized that we may risk losing our clients if we didn’t improve our cloud infrastructure and start focusing on production-grade reliability.”

Bar Fingerman, Head of Engineering

What we’ve done

The Solution

Initially, the team wasn’t sure about outsourcing their infrastructure work.

The engineering team considered hiring an experienced DevOps engineer with Kubernetes skills in-house. However, they didn’t have enough tasks in the backlog to justify a full-time position.

Bringing in someone less experienced wasn’t an option either since a Junior DevOps engineer wouldn’t be able to complete the migration to Kubernetes within a reasonable time frame, nor would they have the support required to make the correct architecture-related decisions.

The team at Bria decided to go with Opsfleet’s packaged services bundle. The bundle included part-time engineering capacity to drive the Kubernetes project and other initiatives forward, along with project management and advisory services provided by Opsfleet’s management team.

“It was straightforward to start working with Opsfleet. We received value from day one, didn’t need a complicated onboarding process, and felt that tasks were moving smoothly and quickly.”

Bar Fingerman, Head of Engineering


The Outcome

By collaborating with Opsfleet, Bria’s engineering team managed to achieve many improvements in their cloud environments.

The introduction of autoscaled Kubernetes infrastructure, monitoring open source tools, and on-call practices allowed Bria to meet the required high availability metrics set by the team to support enterprise clients. The path from zero to production was quick and enabled the team to jumpstart the move to Kubernetes while providing benefits almost immediately.

Regarding cost management and efficiency, the team introduced spot instances to reduce spending in non-critical pre-production environments and helped improve Bria’s training lab by developing a mechanism that queried the AWS APIs to find regions with the most available GPUs.

The introduction of infrastructure-as-code as the main driver for infrastructure changes significantly increased the engineering team’s velocity by making production more stable and predictable and by decreasing variability between pre-production environments.

“Our move to Infra-as-code improved how we work. It’s much easier now to make changes once all our infrastructure is in code, and we have multiple environments that all look the same. All our engineers learned Terraform, which most of them didn’t know before, and now actively make necessary changes themselves”

Bar Fingerman, Head of Engineering

Although Opsfleet assigned one part-time engineer to drive most of the technical work, the availability of domain experts within the Opsfleet organization helped the assigned engineer provide quality results much faster, rather than reinventing the wheel and spending time on learning.

“We love working with Claudio, and although he’s very experienced when he doesn’t know something or isn’t sure about the best path forward, he used the ecosystem of friends inside Opsfleet to brainstorm solutions for us and gain knowledge fast. That’s a huge bonus, and I’m sure we benefited greatly.”

Bar Fingerman, Head of Engineering

Overall, Bria would highly recommend working with Opsfleet to other startups facing similar infrastructure challenges. They appreciated the expertise and value delivered by the Opsfleet team and achieved significant improvements quickly.

“We are very satisfied with the work so far and feel that we’ve achieved a high availability and stability in our systems, allowing us to focus more on cost and general operational efficiency going forward.”

Bar Fingerman, Head of Engineering

“We are very satisfied with the work so far and feel that we’ve achieved a high availability and stability in our systems, allowing us to focus more on cost and general operational efficiency going forward.”

Bar Fingerman
Head of Engineering
quote decorative illustration

Next Projects

Opsfleet Helps Axis Security Overcome DevOps Hiring Challenges and Delivers Infrastructure Success

Learn how Axis Security, a growing SaaS startup with a mix of cloud and on-prem components, overcame the challenge of hiring experienced DevOps engineers by working with Opsfleet.

How Kubernetes Rescued Avantis' Uprise Project and Increased Productivity by 30%

Avantis, a technology company that provides monetization infrastructure for software developers, invested in an internal innovation project, Uprise, to build a user-friendly, riskless, and performance-based advertising product for businesses.

Opsfleet Helps BigPanda Migrate to Kubernetes and Simplify Their Development Process

The case study highlights the challenges BigPanda faced when attempting to make the switch on their own, why they chose Opsfleet, and how Opsfleet improved their development process and leveled up their expertise in Kubernetes.