Codium Case Study
CodiumAI is a growing startup that provides tools for developers to test and analyze their code.
Let's Talk
CodiumAI is a growing startup that provides tools for developers to test and analyze their code.
Let's TalkCodiumAI is a growing startup that provides tools for developers to test and analyze their code. By providing AI-powered interactive code integrity tools, Codium helps developers eradicate bugs and code with confidence.
Codium was using the Google Cloud Platform (GCP) for all their research-related workloads and were worried about being pigeon-holed to one provider. They were motivated to migrate to AWS for the following reasons:
With more experience in Google Cloud, the Codium team were apprehensive about how easy it would be to work with AWS on a daily basis and how the move would affect their developers efficiency. They needed to know that they’d be able to connect to the GPU resources properly with a stable networking solution—from both their office and other private networks used by the team.
Codium’s ML Engineering Architect, Ori Kotek, explains the value of working with a dedicated DevOps expert with vast experience in AWS:
"It’s notable that Roman brought to the table a deep understanding and depth of knowledge in the AWS cloud. I asked ad hoc questions on a daily basis, and Roman managed to solve all my problems, even those I didn’t find a solution for anywhere else." - Ori Kotek, ML Engineering Architect, Codium
Having worked with us previously—using our DevOps as a Service package—Codium were extremely happy with the outcomes. Following this experience, plus a referral, Codium knew that Opsfleet was the team they wanted to work with on their AWS migration.
"Our team was already collaborating happily with Opsfleet, using their DevOps-as-a-Service offering to support development activities. Knowing of Opsfleet’s experience with migration projects, and that they could support multiple clouds (including AWS), it was clear to us why they had been recommended by the AWS team." - Ori Kotek, ML Engineering Architect, Codium
To get things started and secure the funding from AWS, we created a full migration plan, architecture design and figured out the necessary cost calculations. We then worked with the AWS team to approve the project, before kicking it off.
Within just two weeks, we migrated Codium’s research environment from Google Cloud to AWS, making sure not to affect their current on–going development projects or the clients they had on Google. We provided cloud agnostic solutions by relying less on GCP or AWS specific services where possible. The IaaC code was written using the same conventions (naming, variables), and resides in the same repository with similar structure to how GCP code is managed.
In a short amount of time, we had implemented and escorted Codium through the entire migration process, which also included:
"Roman’s depth of knowledge, and the overall Opsfleet team’s support in the background, assisted the project to move very quickly—we completed the migration of most of the research resources to AWS within two weeks!" - Ori Kotek, ML Engineering Architect, Codium
With Opsfleet's support, Codium were able to move away from relying completely on Google cloud to achieve a multi cloud infrastructure. The migration was quick with minimal disruption to the daily activities of the development team or for the end users. The Codium team is now ready to move forward with the second phase of this project, and with Opsfleet's assistance, start working on their DR plans.
"With Opsfleet’s support, Codium are happy to move forward to the second phase of the project and start working on our DR plans, which is another important milestone we would like to achieve in the upcoming months." - Ori Kotek, ML Engineering Architect, Codium
We are very happy with the GPU accessibility on AWS and the stability, and although we are still learning to use AWS in the most convenient and efficient way for us, we are happy with the speed of the migration and the quick results for our team.
Optibus, a SaaS platform that plans and schedules mass-transportation through AI and optimization algorithms, faced challenges in scaling their CPU-intensive system due to cumbersome scaling cycles and custom Python code.
Codium was using the Google Cloud Platform (GCP) for all their research-related workloads and were worried about being pigeon-holed to one provider. They were motivated to migrate to AWS.
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.
