Who is Gingalab?
Founded in 2015, Gingalab is a French company that deploys and automates individualized video content creation, within chatbots, advertising, and community management environments. GingaLab offers marketing technology and video mixing solutions for producers, broadcasters, and brands in Europe and around the world.
Located in the 18th arrondissement of Paris, the company currently employs more than 10 employees.
Gingalab offers a large panel of SaaS solutions. However, its flagship service is called Just Edit. Just Edit is a solution for instant editing of audiovisual content for Social TV. With this solution, you can capture events, interviews, or product demos with a phone or a camera and upload your media to Just Edit. You can also access the Shutterstock library, import media directly, and enjoy automatically generated and high-quality translations and subtitles. Last but not least, you can contact at any moment the international network of professionals from Just Edit Studio to improve your videos.
On the client-side, Just Edit uses a React frontend that calls their APIs directly. Users can authenticate themselves using either their account or their Google account with OAuth2. Once they are connected, they can easily upload the videos they will work on, following a complete video pipeline.
From a Technical Point of View
For their backend, Just Edit uses Django and Django REST framework. Once a video is uploaded, a list of Celery tasks is triggered to create new assets (thumbnails, reencoded videos, etc.). These tasks are performed asynchronously by a set of workers and mostly rely on FFmpeg. Then a set of microservices provides subtitle services on the video by using different well-known SpeechToText providers. In addition, to provide these treatments, their backend also handles the billing of their platform usage.
Gingalab has been using Scaleway Instances from the beginning. They started with the historical Bare-Metal offers C1 and C2. From there, they migrated progressively to virtual instances such as DEV and GP instances. As of today, the company uses the whole Scaleway Elements ecosystem with a large number of Virtual Compute instances, Managed Databases, Managed Load Balancers, Container Registry, and several object storage buckets. They can reproduce the production environment in a staging environment using DEV instances instead of their GP instances.
The company developed their whole Pub-Sub system internally and is running it using Kubernetes Kapsule. The container images used by their solution are stored on Scaleway’s managed Container Registry service. All the data persistence is done in a managed PostgreSQL database. Videos and other media objects are uploaded and stored automatically in object storage buckets.
The managed database is prompted by the different Django backends to perform their stateful operations. They can take advantage of the scheduled backup and restore features that are provided by Scaleway managed databases.
“We chose Scaleway because of the very competitive price and the geographical location of our data. We want to have our infrastructures in France or at least in Europe. Scaleway is a trusted ally in that sense” explains Brice Parent, CTO.
In addition, Gingalab is pleased with Scaleway’s assistance, finding them to react swiftly and efficiently when two small hardware failures were encountered. Brice adds, “we love the console, it is very simple and easy to use.”
As a next step, the company would like to test GPU instances based on Nvidia Tesla chips now that native support for GPU acceleration is provided.
Brice Parent also adds that they are interested in looking into Scaleway Serverless, which could be the right fit for the batch processing of video files.