How a Young Huawei Engineer Saved Millions with Cloud Resource Scheduling Optimization

In August 2019, 25‑year‑old Tong Hao joined Huawei. Two years later, the universal Yao Guang resource re‑scheduling solver he helped develop saved Huawei Cloud thousands of yuan each year, becoming his most rewarding achievement.

Described as the “Yao Guang Youth,” Tong remains modest, emphasizing his love for code and willingness to tackle new challenges.

Born in a small city, Tong’s fascination with computers began in elementary school through a computer club. He excelled in national programming contests, winning the ICPC Asia championship, 13th place in the world finals, second place in IEEEXtreme, and other honors, becoming the most decorated member of his university’s ACM team.

After graduating from a top Chinese engineering university, he chose Huawei over other offers, attracted by the company’s growth opportunities.

His first project at Huawei Cloud focused on intelligent scheduling and operations optimization. The core problem was to intelligently consolidate fragmented physical resources in data‑center pools caused by dynamic virtual‑machine allocation, a situation he likened to “Tetris holes.”

The Yao Guang solver fills these gaps by decoupling underlying algorithms from business constraints and objectives through an intermediate model, allowing users to define custom optimization goals without modifying core algorithms. Deployed in Huawei Cloud data centers, the solver now saves the company millions of yuan annually.

Following this success, Tong moved to cloud identity authentication and access control, aiming to enhance data security by offering richer attribute‑based access control capabilities.

He attributes his growth to a passion for challenging problems, a relentless pursuit of excellence, and the supportive, expert environment at Huawei Cloud.

Looking ahead, Tong plans to continue deepening his expertise in cloud computing, focusing on intelligent scheduling, operations optimization, and security control, while exploring unexplored technical frontiers.