A Year of Video‑Quality Engineering and Growth at Taobao Live

This article is the seventh entry of the "Long‑termism Starts with Small Things" engineering growth series launched by Taobao Tech. It collects sincere reflections from several engineers covering front‑end, back‑end, data algorithms and technical quality.

The author, Xiaocen, holds a Ph.D. in video‑stabilization algorithms and has worked at Huawei, Alibaba, and now Taobao Live, shifting from pure algorithm research to video‑quality and experience improvement for live streaming and short video.

Key experiences in the past year:

Developed an automated attribution program that identified low‑quality causes for thousands of streamers and pushed personalized messages, proving that technical initiative can directly drive results.

Implemented a real‑time low‑quality detection feature for live rooms, completing the full product cycle from requirement discussion to deployment and observing tangible impact.

Engaged closely with streamers, providing equipment recommendations and troubleshooting, which built trust and yielded valuable feedback for the product.

Acted as a demand owner, consolidating streamer complaints and suggesting improvements to beautification, green‑screen, and artifact‑reduction algorithms.

The Taobao audio‑video technology team is responsible for high‑quality, low‑latency video solutions (S265 encoder, STaoVideo enhancement, MD‑VQA quality model, TMPS media processing, GRTN transmission). Their work supports the broader content strategy and has earned recognition in international competitions.

In summary, the author reflects on the significance of the transition from pure algorithm work to product‑oriented engineering, emphasizing the importance of delivering business value, cross‑domain communication, and continuous improvement of video quality.