Tu Changyang, Vice President of Technology of Kechuang and Wisdom Buds: The road of technology is the road of challenges, and I don’t want to waste time in big factories
Tu Changyang, Vice President of Wisdom Bud Technology, has more than 20 years of experience in the high-tech and Internet industries, and has worked in companies such as Cisco and WebEx. He has been engaged in providing technical operation services of collaborative open cloud platform for enterprise users for a long time, and is responsible for the technical operation of cloud platform and the development of related cloud management software. Wisdom Buds was established in 2007 with global headquarters in Suzhou and branches in China, Singapore, England, Canada and Japan. Tu Changyang is in charge of 700+ R&D teams around the world, managing data, development, artificial intelligence, etc.
Text | babayage
Edit | Laugh Laugh
The awakening moment of technicians recalls his more than 20 years of technical life, and Simon Tu Changyang marked several key points with "wake up-time":
Awakening moment 1. Computers can do business: Tu Changyang first came into contact with computers in high school. He wrote and drew in Basic language on Apple Macintosh. "It doesn't feel like a small game." It was not until he went to university and used computers to perform mechanical calculations that Tu Changyang realized the great help of computers for scientific computing.
Awakening Moment 2. Magical computer hardware: The computers in the computer room often broke down during the university days. Most of the reasons were the failure of the network connection to connect with the host computer. Forcing Tu Changyang to read the instructions to understand the principles of various computer hardware devices. He can still remember the shock when he saw the host for the first time. "It's a big machine with a bunch of disks in it, and there are old machines with various punch cards in the corners."
Awakening Moment 3. The charm of networking: After graduating from university, after a year of work experience within the system, Tu Changyang completely understood one thing: he only wants to do challenging and difficult things, and never wants to live a peaceful life. He joined Founder of Peking University in 1999, and had the opportunity to truly understand the IT industry. He learned that the computing power of computers connected to the Internet can make computers hundreds of times more efficient, which can be regarded as a foundation for future engagement in the SaaS industry.
"I didn't have lofty ideals at that time, but I knew that I was unwilling to do simple things, and liked challenging and difficult technical careers." To this day, Tu Changyang has continued this principle of life decision-making. It is precisely because of this that in 1998, the Internet bubble of "simple technology but very profitable" once made him deeply confused.
I can't understand the "too simple" Internet bubble, challenge myself to join WebEx
In 1998, any dormitory building for science and engineering majors in any university in Beijing had more than one xx.com entrepreneurial project. Many colleagues around Tu Changyang also chose to change jobs and join various ".com" companies.
Tu Changyang used "confused" to describe his mood at the time, "I really couldn't understand it for a year or two, what technical content does writing a web page have? But the market demand is very large, the development is good, the financing is fast, and the salary is also high. The pursuit of technological advancement seems to be of little value, and I am indeed a little confused.” It's not that he's completely unmoved, but in any case, it's hard to convince himself to do something "too easy".
In 1999, in the department where "people run wild", Tu Changyang, who was stuck in the same place because of confusion, became a more reliable person in the eyes of the boss, and he was given a team to be responsible for the development of the communication module of the printing equipment. . This technically challenging job calmed Tu Changyang's emotions. At that time, he basically didn't understand management. Like many technical gurus who have just transferred to management positions, he only relied on "You can do what you can do, but I will do it if you can't do it." "The leadership of the team has brought the small team full of energy and smooth delivery, and also initially recognized the difference between "team" and "individual".
In 2001, the bang of the global Internet bubble burst spread to China's Internet industry, and companies were closing down every day. On one side, there are many sorrows and sorrows of Internet people, and on the other side, Tu Changyang, who has always adhered to the technical challenge route, got a job offer from WebEx in the United States. WebEx, the world's number one online conference provider, provided him with a job that completely suits his taste. Work: Difficult, challenging, and clear.
Technology needs to serve customer scenarios. It is easy to hear it, but it is wrong to do it in WebEx. Tu Changyang has completed a breakthrough growth in his career. At this stage, he also marked two awakening moments:
Awakening Moment 4. Growth of technical capabilities: At the beginning of the 20th century, there was a significant gap in the technical strength of IT engineers between China and the United States. His work experience in WebEx multinational teams and his experience in researching and developing products for global customers allowed Tu Changyang to gradually grow in both technology and management. stand up.
The lessons of Awakening Moment 5. Blood, I understand the true meaning of technical services and customers:
"Technology serves customers. It's easy to understand when you hear it, but it's wrong when you practice it." What really made Tu Changyang unforgettable about this sentence was: a super-large customer used WebEx products to hold an online board of directors, and suddenly a malfunction occurred. The boss was very annoyed and ordered everyone to solve the customer's problem immediately. Netspeed's top technical brains, including Tu Changyang, got together, and it took 12 hours to lock the problem and repair it. The team was smiling and ready to ask the boss for credit, but the boss was waiting with the most violent firepower ever. Criticism of - "For 12 hours, a group of people were thinking about solving technical problems, but no one thought about doing customer service first, and eventually lost this customer, which was very important to WebEx at the time. This lesson is particularly painful. , so that I suddenly became enlightened after that, reminding myself to think about what services and problems I can provide to customers.”
In addition, it is worth mentioning that Tu Changyang did not realize that he became a pioneer explorer of the business model of To B SaaS: WebEx is one of the two earliest start-ups in the SaaS model (the other is Salesforce), Tu Changyang The big things and small things of the SaaS model can be called the door clear.
Merging into Cisco has officially become the most headache for managers to lay off employees, and the most anxious large enterprise disease
In 2007, Tu Changyang, who was on a business trip to the United States, met the boss in the elevator. This tough, unsmiling boss actually greeted him with an unprecedented smile, "I thought at the time, this boss used to be a VP-level person and said. Just fry it, why did it suddenly become nice?"
On May 25, 2007, the news of Cisco's $3.2 billion acquisition of WebEx was officially announced. The mystery of the boss's good mood was revealed, and Tu Changyang's 11-year Cisco career was officially announced.
(In October 2007, Wisdom Bud PatSnap was established.)
Before WebEx was acquired by Cisco, the typical Silicon Valley startup culture was simple, straightforward, technology and product first. In Cisco, there is a big difference, requiring formal management training, implementing rigorous process systems, and experiencing a long period of running-in and adaptation. Because of his strong adaptability and good understanding of management, Tu Changyang has gradually grown into a senior technical leader. At most, he managed more than 300 engineers distributed all over the world.
Tu Changyang · Q&A on R&D management
Q1. The most troublesome management problem?
A. The first layoff. After the 2008 economic crisis, the company decided to optimize the R&D personnel of the whole company (at that time, the global IT companies were laying off staff), and a fixed 10% layoff was required. At that time, the team was either recruited by me or worked together for many years. Colleagues, and our BU's performance and growth are very good, the 10% layoff ratio is really a headache.
Q2. How to finalize the 10% list?
A. My idea is to first determine who must stay, including the core of the team, potential employees, and performance in previous years. The final list will be evaluated based on performance.
Q3. R&D performance is a common problem in R&D management. What performance evaluation method do you use?
A. First of all, it will not be a single indicator evaluation. Technical people are too smart, and he will always have a way to deal with the single indicator you set.
The starting point of R&D performance management is recruitment, and it is necessary to recruit excellent people from the beginning of recruitment; the second is the matching of people and positions, people must be placed in the right position, and emotions and sense of achievement must be taken care of; third, subjective evaluation is a must in the software industry. Indispensable, probably more than 50%.
Q4. The subjective evaluation rights of large teams must be decentralized. How to ensure the relative consistency of subjective evaluation standards?
A. First of all, based on the hierarchical authorization based on the trust mechanism, the top leader must not be able to control everyone, but there must be time to maintain communication with the middle management and align their ideas; secondly, the reward mechanism is set. What we do is to select star engineers and clarify what we will reward. Behavior, characteristics and evaluation dimensions, such as cross-business collaboration, such as technological innovation, pick out outstanding behaviors and results, reward each quarter, tell everyone which direction should be developed, and further build a general consensus through individual honors.
The distribution of talents is basically normal distribution, you only need to grasp the head and tail.
After reviewing the biggest impression of his 11-year career at Cisco, Tu Changyang summed up the gain and a headache: the gain is that he has personally experienced how to achieve excellence in an enterprise that pursues excellence. Cisco is a company that guarantees its competitiveness through a large number of mergers and acquisitions and abolitions. In any business field, Cisco only accepts the top 2 market position, and if it does not meet the standards, it will give up decisively. Such an air."
The headache is that a lot of working time in large companies is spent on communication, and the most tormented part of Tu Changyang is unnecessary repeated communication. "After 2015, I felt that a lot of my time was wasted. I kept meeting, showing my team, and coordinating with other multinational teams, which led to slow product innovation. For example, Zoom was acquired after WebEx was acquired, and innovation changed. The slow timing developed rapidly... I finally decided to leave Cisco, and it was four of my bosses in one and a half years. In front of the new bosses, I introduced what our team does, what strengths we have, and what we plan to do in the future. What, said the same thing four times, that was the straw that broke the camel's back."
In 2018, Tu Changyang left Cisco and joined the domestic technology innovation intelligence SaaS service provider Wisdom Bud.
Big data artificial intelligence helps "buds" grow
In 2017, Zhang Jihui, the founder of Wisdom Buds, came to visit Tu Changyang for the first time, but the latter declined the warm invitation, "I couldn't understand the value of patents. to the other side".
Soon after, Zhang Jihui visited again, described in detail the mechanism of patent data accumulation in driving the direction of enterprise innovation, and described a grand blueprint for the R&D and innovation of SaaS. This vision touched Tu Changyang, "When I got in touch with the team, I was full of vigor. But it's a bit inexperienced, and it reminds me of myself when I first joined WebEx."
"After so many years, I haven't changed. I still can't talk about how grand the ideal is, but I still want to try to challenge difficult things." Joining Wisdom Bud, Tu Changyang's technical strategy and leadership are beyond doubt, but the technical details have a certain angle with what he is good at. Deep understanding.
Kechuangren: Can you introduce the business strategic plan of Smart Buds and the technical realization path you have outlined?
Tu Changyang: Wisdom Buds first entered the market with PatSnap global patent database products, and its current strategic direction is a SaaS service provider of technological innovation intelligence. The product matrix includes: PatSnap global patent database, Innosnap intellectual property management system, Insights patent analysis system, Discovery Innovation information system, Life Science series databases, etc., serve different cycles, different links and different roles of innovation and R&D.
In terms of data sources, Wisdom Buds has accumulated more than 170 million global patent data, more than 220 million chemical structure data, as well as comprehensive and massive data such as tens of millions of financial news, scientific literature, market reports, and investment information. New materials, automobiles and other industries establish vertical industry databases. In order to achieve product-level applications, artificial intelligence technologies such as machine learning, computer vision, and natural language processing (NLP) must be inseparable.
In terms of data processing, through the support of artificial intelligence technologies such as NLP (natural language processing), CV (visual computing), and machine learning, Wisdom Buds can realize automatic and intelligent data mining and text mining, and achieve deep semantic search . Up to now, Wisdom Buds can provide 3 language retrieval methods, patent data is updated every week, 90% of the data can be processed automatically by machines, and the product response speed can be achieved in seconds.
Full-link data capabilities and cross-domain replication capabilities are equally important. Wisdom Buds has formed a powerful "data processing factory" model, tempering data capabilities in all aspects: First, core data processing capabilities include image processing capabilities, natural language processing, optical Character recognition (OCR) and machine translation, etc., ensure the comprehensiveness and timeliness of data; secondly, in data processing, Wisdom Buds has a unique advantage - an integrated system of data annotation engineering and data feature engineering, which realizes intelligent data annotation It can greatly improve the processing efficiency, and also realize the self-circulation of training samples, which greatly reduces the processing cost. Third, the intelligent matching and classification system based on machine learning has a classification accuracy rate greater than 96.3%. At present, the Wisdom Buds data processing factory has entered the 2.0 stage. In terms of human-machine integration, automation 2.0, scalable and replicable, and cross-domain algorithm-driven and knowledge link capabilities, Wisdom Buds is still improving.
Wisdom Buds has invested a lot of energy in the basic research of AI, and has developed a pre-training model for specific industries with its own core technology, and has achieved good application results. Through the application of cutting-edge artificial intelligence technologies such as deep learning, natural language processing, and pre-trained language models, automated and intelligent data analysis and text mining in massive global multilingual patent texts are realized, and deep semantic analysis is further realized. Provide more accurate semantic retrieval services.
Starting from different data sources, Wisdom Buds conducts data governance in fine-grained text entity extraction, entity association, problem association, domain association, feature selection, etc., through domain semantic map, text/image training samples and pre-training models, etc. , Through the self-developed machine translation engine system, Wisdom Buds has broken through the language barriers between different patent texts, making it possible to analyze, retrieve and semantically mine TB-level patent texts around the world.
The patent intelligent semantic retrieval system is mainly based on the patent pre-trained language model, and uses the BERT twin network to obtain the shallow semantic information of the text based on the DSSM architecture. At the retrieval level, it adopts a combination of traditional keyword retrieval and vector retrieval. It not only ensures the coverage of retrieval results, but also ensures the relevance of retrieval results. Through the application of the intelligent semantic sorting method based on the pre-trained language model, the patent intelligent semantic retrieval system of Zhiya Bud has achieved great technological breakthrough and commercial value.
Kechuang people: According to your understanding, what is the value that Wisdom Buds will truly realize for customers in the future?
Tu Changyang: An easy-to-use, extensive database, a wide range of intellectual property products, and a forward-looking innovative R&D system, of which I am most interested in the innovative R&D system. The essence of serving the development of intellectual property is to serve the development of innovation. Based on patent big data, Wisdom Buds not only continue to serve intellectual property and innovation protection, but also serve the ecology around "innovation" - such as our innovation intelligence system Discovery that serves R&D decisions , the scientific and technological innovation evaluation system TFFI serving financial institutions, the Pharmsnap system serving the research and development of innovative drugs, etc.
The real value realized by Wisdom Buds for customers is to provide efficient software and tools for the whole process of innovation with enterprises as the main body, so as to help the efficiency and high quality of innovative R&D activities.