How we delivered product work inside Kaltura's live Cloud TV platform
For more than three years, a dedicated Aetsoft team developed, tested and supported product changes for Kaltura's enterprise customers, including work on its AI recommendation engine.
Kaltura's Cloud TV platform supports television services for telecom companies, broadcasters and media businesses. From 2019 to 2022, a dedicated Aetsoft team worked inside that platform on product changes for Kaltura's enterprise customers. Part of the engagement was engineering work on Kaltura's AI recommendation engine.
001
ClientKaltura, enterprise video and Cloud TV
002
ChallengeDeliver product changes inside a platform already serving viewers
003
SolutionA dedicated Aetsoft team combining development, quality assurance and ongoing support
Engagement lead
Pavel Sivayeu
CTO, Aetsoft Inc
Ran delivery and led the Aetsoft team across the engagement, 2019 to 2022
The platform was already on air
Kaltura's Cloud TV runs live and linear channels, video on demand and cloud DVR, delivered to phones, browsers and smart TVs. Kaltura reports a carrier-grade SLA of 99.995 percent, across 15 or more global television services with tens of millions of viewers between them. People expect television to be there at any hour, in every region the service sells into.
That sets the terms for every change. Code goes into a system that is serving people while it is being deployed. It reaches device fleets already sitting in living rooms, under the security expectations telecom companies and broadcasters bring with them. A defect reaches somebody's television before it reaches a support ticket.
Kaltura needed engineers who could work at that standard on its enterprise customer work, and keep working at it for years.
Kaltura needed engineers who could work at that standard on its enterprise customer work, and keep working at it for years.
How the team worked
Development and quality assurance
The team worked in a product environment built on .NET and C#, running on AWS, with content management frameworks around it.
The product around that environment is wide. It covers the following areas:
- media preparation running from content acquisition through transcoding and packaging to encryption and delivery to the CDN
- business logic for linear, live and on-demand content across subscription, transactional and advertising models
- cross-device DRM with on-the-fly encryption, supporting both online and offline playback
- a player Kaltura reports as supporting more than 8,000 device types
A change in one part of that environment can affect workflows and customers elsewhere in the product. Working in it started with reading. The code was years old, several teams depended on it, and a change that looks local in a platform this size usually reaches further than the file it sits in.
Manual and automated quality assurance ran inside the team, alongside development, rather than as a stage handed to somebody else once the code was written.
Delivery and the recommendation engine
Delivery for Kaltura's enterprise customers
Work was planned and estimated in sprints, with the backlog groomed as new requests arrived. Coordination ran daily inside the team and weekly with Kaltura, tracked in Jira, with reporting on a fixed cadence.
Requirements kept moving while the work was under way, which is normal when the customer is a large television business with commercial commitments of its own. Short cycles and honest estimates are what let a team absorb that inside the plan.
The team stayed with Kaltura's enterprise customers across the whole journey, from onboarding through ongoing support. Estimates, trade-offs and technical constraints had to be explained in terms those stakeholders could act on.
Engineering on the AI recommendation engine
Part of the engagement was engineering work on Kaltura's AI recommendation engine.
Recommendations decide what a television service looks like when it comes on. A large catalogue pays for itself through what viewers find in it. So the recommendation layer sits between the content a business has licensed and the viewer deciding whether to keep subscribing.
The personalisation in Kaltura’s products is built-in analytics and AI-driven metadata delivering tailored recommendations and targeted content, running alongside advanced and unified search.
For a broader explanation of how systems like this are designed, see our guide to AI recommendation systems.
Results
“What stood out most was their strong sense of ownership and accountability. They approached the work as true partners rather than just a service provider.”
Alex Glushenkov 
former SVP Engineering at Kaltura Verified Clutch review.
After more than three years, Kaltura transitioned the team's scope to its internal engineers through a planned handover.
Across the engagement, Aetsoft carried the work from a customer requirement to a tested change, and then supported it in the product afterwards.
Aetsoft's close work with Kaltura's enterprise customers strengthened those relationships and helped secure a meaningful share of the annual recurring revenue (ARR) they bring in.
Frequently asked questions about Kaltura's live Cloud TV platform
-
What did Aetsoft do for Kaltura?
From 2019 to 2022, a dedicated Aetsoft team developed, tested and supported product changes for Kaltura’s enterprise customers. The team worked with .NET, C# and AWS, handled manual and automated QA, and contributed to Kaltura’s AI recommendation engine.
-
What is difficult about developing for a live Cloud TV platform?
Changes must be released while the service is running across regions and thousands of devices. Even a small update can affect other parts of the platform, so testing and release control are critical.
-
How was the team managed?
Aetsoft ran delivery for the team: sprint planning with estimates, backlog grooming, daily coordination inside the team, weekly sessions with Kaltura, and regular reporting, tracked in Jira alongside Kaltura’s own work.
Contact us
Shoot us an email with your request, and we will contact you within one business day.