Details
DreamWorks feeds metadata about compute jobs, scene complexity, hours invested in assets, and storage behavior into a big data pipeline. Machine learning algorithms then establish a baseline of normal behavior by running synthetic transactions that replicate artists' workflows, and flag anomalies — for example, unusual latency on a file endpoint — so engineers can intervene before artists are affected. DreamWorks CTO Jeff Wike has publicly described the system as directly tied to the studio's ability to make films on schedule.
Have evidence about DreamWorks Animation's AI practices? Submit a report.
Report a Sighting →