
A self-directed port team took a PlayStation VR comedy adventure to Xbox One and Nintendo Switch — and delivered five months ahead of schedule.

Partner studio: Puny Human
A comedy action-adventure played from the perspective of a seated “chairorpian,” directing the title character through a bizarre, densely voiced universe. Originally built for PlayStation 4 and PlayStation VR with a high-fidelity art package, later extended with free downloadable content.
Following a well-received PlayStation and PC launch, the partner wanted the title on Nintendo Switch and Xbox One, packaged with the free DLC. Switch was the binding constraint: the game was simply too large to transfer directly without modification, and reducing it had to be done without visibly compromising the artistic vision. Xbox One brought its own problems, including account-swap and save-data failures that returned players to the main menu or crashed the game. The partner retained marketing, store management, localization and art, and needed the port team to run itself.
A dedicated team of engineers, QA testers and producers operated separately from the client's development team for the first time — the client set deadlines, the team managed the path to them. Engineers split across the two platforms and defined their own code review process.
Optimization was approached as a data problem. Pre-programmed agents built on the Gauntlet automation tool ran continuous performance capture, including agents that warped repeatedly around the game to gather metrics unattended. Texture and lighting audits identified assets carrying unnecessary high-resolution textures or dynamic lighting in places players would never notice, so reductions could be targeted where they would not be seen.
Custom code was written for account and user management to resolve the Xbox One save and profile failures. The underlying Unreal Engine 4 version was upgraded several times mid-project; every resulting defect was regression-tested across platforms and against the shipped build, with automation absorbing the routine load so testers could concentrate on engine-upgrade regressions.
Engine: Unreal Engine 4.



Describe the problem and we will scope the smallest useful engagement.