decorative pattern
decorative pattern

DevOps as a Service Success Story

Imagen AI deliver DevOps projects faster and at the highest engineering standard for over 3 years

Let's Talk
Imagen AI

The Client

Imagen is an AI editing platform that saves professional photographers time by building a custom AI model that automatically edits their images based on their unique style. They provide photographers with the fastest way to batch edit their photos, allowing them to concentrate on their creative work rather than the tedious editing process.

Region
Tel Aviv, Israel
Industry
Software Development
Main Technologies
AWS, Kubernets, GPU
Services
Date of Project
Problem

The Challenge

When Shachar Polak, head of engineering, joined the company in its early stages, the first order of business was to build and prepare its cloud infrastructure for scale. 

As a B2C company, they were expecting massive growth in their user base and had to ensure their infrastructure could facilitate that scale without impacting SLA. 

Imagen was looking to modernize and standardize their cloud infrastructure because: 

  • Their platform was built on a traditional application architecture, which wasn’t ideal for scale. 
  • They used a self-hosted database that required constant manual maintenance, monitoring, and right-sizing.
  • Their database was running on instances and was not a managed service.
  • They had no proper CI/CD pipelines to facilitate faster, more structured, and safer deployments. 

Typically, with many urgent DevOps projects to deliver, any engineering manager would hire an in-house DevOps engineer. 

But for Imagen AI, it wasn’t a fast and scalable solution:

  • The variety of infrastructure challenges required different sets of skills and expertise, which a single engineer can’t possibly hold. 
  • They needed the flexibility of having more than one engineer to facilitate peaks.  
  • They had to onboard someone fast and jumpstart urgent projects. 
  • They didn’t have the time and resources to start a lengthy hiring process. 

ImagenAI knew that outsourcing their DevOps projects to a professional agency was the only way to get started immediately, with minimal hassle.  

After meeting with many DevOps services agencies, Shahar chose Opsfleet as the partner to help ImagenAI design and build its cloud environments.   

“I was looking for a long-term partner I could trust, professionally devoted, and with whom I would enjoy working. And Opsfleet was the only company that ticked all my boxes.” - Shahar Polak, Head of Engineering, Imagen AI
What we’ve done

The Solution

ImagenAI partnered with Opsfleet to support their cloud journey and hired a DevOps specialist for a monthly ‘DevOps as-a-service’ model.

From day 1, the Opsfleet management team was heavily involved in the design and architecture. 

Opsfleet’s DevOps specialist was onboarded in a few days, learned about their business and application, and immediately started tackling the most urgent projects. 

“From the first meeting, I felt these guys would help me deliver my projects faster and to the highest standard.” - Shahar Polak, Head of Engineering, Imagen AI

Opsfleet helped ImagenAI build the cloud infrastructure that supports and scales its AI-powered platform:

  • They modernized their compute workloads to run on top of Kubernetes to achieve faster and more advanced auto-scaling capabilities.   
  • They shifted their Databases to run on MySQL Aurora, a managed DB service that automatically grows and scales the database to facilitate demand. 
  • Implemented SageMaker ecosystem operations
  • Enhanced their Storage operations - EFS and FSx Lustre setup for ML jobs
  • Optimized ML Job duration for efficiency. 
  • Implemented Custom Metrics and Monitoring for ML Jobs.
  • They designed and built structured CI/CD pipelines for faster, more structured, and safer version deployments.  
  • To comply with their customers' requirements, they also implemented a DRP solution that migrates workloads to a different AWS region in case of an outage.
“Their DevOps specialist has been with us for the past three years, and despite working remotely, he really feels like a part of our team.” - Shahar Polak, Head of Engineering, Imagen AI
Results

The Outcome

For the past three years, Opsfleet’s DevOps specialist has worked full-time at ImagenAI, helping them build and scale a high-standard cloud environment that supports their continuous and massive scale. 

During peak times with many DevOps project deliveries in the pipeline, Imagen AI has the flexibility to onboard one or more additional engineers to facilitate the demand. 

Besides leveraging our valuable technical resources, Imagen also regularly consults with Opsfleet’s DevOps experts regarding their cloud architecture, cost savings, and other strategic initiatives. 

"Knowing that I have flexible access to top DevOps resources whenever I need them is reassuring to an engineering manager because I don’t need to think about the long and exhausting hiring process.” - Shahar Polak, Head of Engineering, Imagen AI

“Opsfleet helps us close DevOps gaps, remove infrastructure bottlenecks, and reduce our time to market.

Their dedication to our success helps us deliver any DevOps project at any scale.”

Shahar Polak
Head of Engineering
quote decorative illustration

Next Projects

Bria.ai migrated its EKS workloads to AWS Graviton and achieved better application performance with lower costs.

Learn how Opsfleet helped the engineer team at Bria.ai to seamlessly shift their EKS applications to run on top of Graviton Instances.

Opsfleet Helps Bria.ai Migrate to Kubernetes and Simplify Their Development Process

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.

How Opsfleet's DevOps-as-a-Service Helped Konnecto Scale and Improve Their Infrastructure

Konnecto, a startup in the consumer intelligence space, experienced a backlog of infrastructure and DevOps tasks as the company and product grew. While considering hiring an in-house DevOps engineer, the team faced challenges finding qualified candidates within budget.