Interview with iQIYI CTO Liu Wenfeng on Cloud Computing, AI, and Multimedia Innovation

In a recent LiveVideoStack interview, Liu Wenfeng, Chief Technology Officer of iQIYI, shared his career journey and current focus areas. He discussed his previous roles at Intel China and VMware China, emphasizing his long‑standing interest in system‑level software, Linux kernel research, and the evolution from hardware‑centric work to cloud and application layers.

Liu explained how iQIYI’s cloud platform was built to support rapid growth, providing a stable, efficient, and highly scalable infrastructure. He highlighted the company’s efforts in big data, AI, and comprehensive technical product development.

The interview also covered work‑life balance, with Liu noting the demanding schedule of internet entrepreneurs and the importance of family support.

Liu recommended the book “The Pixar Playbook” (Lawrence Levy) and related works, drawing parallels between Pixar’s blend of technology and creativity and iQIYI’s own innovation strategy.

Regarding user experience, iQIYI has launched 8K VR, Dolby Vision, HDR10, and Dolby Atmos to deliver immersive video and audio. To reduce bandwidth costs, the company optimizes codec algorithms (participating in AVS2 and AV1 development), improves transmission protocols with its self‑developed HCDN technology that combines CDN and P2P, and adopts a hybrid cloud model.

Since 2014, iQIYI has heavily invested in AI, applying it across the entire video lifecycle—from content creation, editing, and restoration to recommendation, distribution, and monetization. AI has enabled efficient video review, automated clip selection for large‑scale productions, and AI‑driven casting tools.

Liu identified future opportunities in AI‑driven multimedia, next‑generation codecs such as AV1 and H.266, and the impact of 5G on bandwidth and immersive experiences like 8K VR.

He confirmed iQIYI’s commitment to open‑source contributions and community engagement, mentioning recent technical talks at conferences, the establishment of a technology brand team, Open Day events, a multimodal video person‑recognition challenge, collaborations with CCF, and the release of nine open‑source projects including the DPVS load balancer on GitHub.