Who can change what our machines and records do

In the spring of 1999 a cylinder-block machining line at Qingtan Diesel Engine Works stopped for two days, and I was the junior engineer sent to find out why. The machine builder's engineer, working from his own office abroad over a dial-up modem, had loaded a program written for another customer's plant. Nobody at Qingtan had known that he could. I spent two nights comparing printouts line by line with an older colleague, and the question I took away has stayed with me for twenty-seven years: who is able to change what our machines, ledgers and records do, and does anyone here know about it?

I grew up in Changzhou, where my father maintained looms at a state textile mill. I studied industrial automation at Yanhu Institute of Technology, graduating in 1998, and joined Qingtan, a state-owned engine maker in the city, as an assistant engineer in its machining workshops. Over seven years I moved from clearing PLC faults on night shifts to leading the controls group for three lines. It was an education in how much of a factory's knowledge lives in people's habits rather than in its documents.

In 2005 I joined Kellmann-Hua Construction Equipment, a Suzhou joint venture between a fictional German excavator maker and a Chinese partner, to help introduce a manufacturing execution system. The German parent had a global template and expected us to adopt it unchanged. Much of it was right. Some of it assumed suppliers, shift patterns and quality checks that Suzhou did not have, and I learnt to tell the difference by standing at the station where a template met a real operator. I also studied part-time for a master's degree in software engineering at Liyuan Graduate School of Software, finishing in 2010, because I wanted to understand the systems I was configuring well enough to argue with their designers.

From 2011 to 2015 I worked on the other side of the table, as delivery director for equipment makers at Zhiluo Software, a Nanjing developer of manufacturing and planning applications. We implemented systems at thirty-one Chinese manufacturers. Delivering software taught me how a vendor's promises are formed, which commitments a delivery team can keep and why projects tend to fail in the third month rather than the twelfth. It also taught me to be specific about domestic software. Some of our modules suited Chinese plants better than anything imported; others were not ready, and I said so to clients more often than my sales colleagues liked.

I joined Ruoshan Heavy Machinery in January 2016 to build systems for its overseas business, first from Singapore and later from Changzhou. Our machines were selling in more than seventy countries, but dealers ordered parts by spreadsheet and sent warranty photographs through messaging apps. In March 2019 I became chief information officer. Since then my work has been shaped by four questions that every Chinese machinery maker with an export business now faces: how to run many plants on common systems, how to serve dealers abroad without moving their customers' data where it should not go, how to answer the policy push for domestic software without breaking production, and how to keep plant networks closed when dozens of outside engineers need to reach our equipment.

I lead 610 people across Changzhou, the plants, Singapore and our regional companies, with seven direct reports. I ask two things of them. The first is to write down how a process actually runs before proposing to change the system that supports it. The second is to tell me early when a cutover, a pilot or a vendor is going wrong. Our Huai'an cutover in 2024 failed partly because a team sensed trouble and waited for certainty before speaking. I have tried since then to make early bad news cheaper to deliver, mostly by thanking the people who bring it.

I report to our president and present to the board, its strategy committee and its audit committee, but I am not a director of Ruoshan. In that seat I have learnt that directors rarely need more technical detail. They need to know which decisions are reversible, which are not and what each will cost if it goes wrong. I am now preparing to serve as an independent director at a company outside our industry, where I hope to ask those questions from the other side of the table.

Inside our industry I am probably known for three positions that have cost money in the short run: that domestic substitution should be sequenced by the maturity of each process, that dealers' customer data should stay in the dealer's region, and that a factory AI pilot should be stopped as readily as it is started. I hold them because I have watched the alternatives being tried, sometimes by me.

Open to Board appointments · Speaking invitations

Languages / Mandarin Chinese · English