In September last year, after Wu Yongming became the new CEO of Alibaba Group, he put forward two strategic priorities: "user-first and AI-driven".
As the first business unit of Ali to implement AI applications, and Wu Yongming's first internal entrepreneurship project in Alibaba, Alimama has attracted special attention.
In the article "The Past of Alimama, When Wu Mama First Started an Internal Business", we review the ups and downs of Alimama's initial period. On this basis, this article will continue to tell how Alimama applied AI technology innovation to advertising recommendation algorithms and ushered in its own heyday.
Only after understanding these glorious journeys of Alimama can we better imagine and look forward to the "Ali in the era of AI e-commerce" that Ma Yun and Wu Ma are looking forward to.
(In the recruitment of the Zuolin Right Fox community, link the front-line industry and investors, welcome group friends who can provide us with new views, new angles and new information, you can add the group owner WeChat (aqingcjx) to apply to join)
Wu Xuejun zigzags "entered Tao" and recruited young talents
In May 2010, Wu Xuejun (Tie Xiang) resigned from Baidu.
At that time, under the influence of Facebook, a wave of social networks was being set off in China, and Renren and Kaixin.com were both rising future stars at that time. Under the situation of cooking oil in a blazing fire, Wu Xuejun can't help but want to throw himself into the wave of social networks to take a look and break through. After several searches, he set his sights on the Tianya community.
In 2010, the Tianya community was in a highlight moment, and a large number of "Tianya Divine Posts" came from this period, but this is already the afterglow of the sunset.
With the storm of the mobile Internet, mobile social media platforms such as Weibo and WeChat were born one after another. The mobile Internet is a deep vertical innovation, and Tianya was still taking a large and comprehensive community model at that time, coupled with internal inaction, relying only on a few moderators with no salary, it was impossible to beat the corporatization of others. It's starting to fall behind.
Only two months after joining Tianya, Wu Xuejun discovered the problem. After gaining an in-depth understanding of the company, he believes that with Tianya's team and corporate culture, he will definitely not be able to achieve his ambitions here.
Wu Xuejun, who was unwilling to waste his years at the end of the world, immediately began to plan the next stop of his life. In fact, before joining Tianya, Wu Xuejun had another choice - Taobao's Liu Zhenfei (Zhenfei) and Mei Jian (Sanduo) once threw an olive branch to him. Wu Xuejun was moved, but he had already accepted Tianya's offer at that time, he didn't want to lose his trust, so he could only sigh "I hate to see each other late", and then declined politely.
After realizing that Tianya was not a good master, Wu Xuejun remembered Liu Zhenfei again. Maybe it's fate, although two months have passed, Taobao's position is still vacant. As a result, Wu Xuejun joined the Taobao Advertising Technology Department smoothly as the head of the algorithm team, reporting to Mei Jian, the technical director of the advertising technology department.
Wu Xuejun
In hindsight, Wu Xuejun's joining Taobao can be said to be a good time. Although the revenue of the ad technology department was still very small at that time, only a few hundred thousand a day, it had already begun to enter the fast lane of growth.
At the end of 2009, the diamond booth, which had been tepid in development, ushered in a revision, and opened an additional resource position in addition to the original sales resources - a position of 300*250 on the third screen of the home page. This is a milestone for the Diamond booth. After entering the homepage, it has enough liquidity, and the whole model begins to have a working basis.
During this period, Taobao's advertising technology department also launched another important product line - Taobao customers, which focus on the CPS model. At that time, the media and the traffic side were keen on CPC marketing, and they generally had a wait-and-see attitude towards the CPS model like Taobaoke. Therefore, Taobao customers almost did not encounter any strong direct opponents in the early days of their birth.
At the same time, for e-commerce sellers, CPS is naturally a low-cost marketing model - whoever can help me find consumers and sell things, I will give the commission to whomever I can.
The traffic side has found a new monetization channel, and Taobao merchants have also found suitable promoters, and Taobao customers want to connect with both ends immediately have a large number of resources pouring in.
Diamond booth (CPM), Taobao customer (CPS), coupled with the earliest through train (CPC) business, the three carriages of Taobao advertising products have taken shape, and they have begun to drive rapid revenue growth.
Not long after Wu Xuejun joined, Taobao's advertising revenue exceeded the 1 million yuan per day mark. To this end, the team also held a very grand celebration banquet.
Before Wu Xuejun, Ali had almost no experienced and influential algorithm talents, so his joining was like a banner that established Ali's appeal to algorithm talents. In addition, Alimama itself has rich data and its revenue is growing rapidly, and its talent attraction is increasing.
So during that period, the team successively recruited a group of young talents such as Gai Kun (Jingshi), Jiang Long (Tanzong), Yuan Quan (Yuan Quan), Hu Yunhua (Wu Gou) and Yan Qiang (Shaocheng). Among them, Gai Kun later became the face of Alimama and was widely known to the outside world.
Rush to the aid of hand-to-hand Tao construction, and the recommended team is divided into two
In 2012, Mei Jian was transferred to take charge of other businesses, and Wu Xuejun took over Taobao's advertising technology department from him. At the end of the same year, Taobao's advertising technology department was re-upgraded to Alimama Division.
At that time, there were three business developments in Taobao's marketing business that were particularly conspicuous: one was Taobao customers, the second was Tanx, and the third was the development of wireless terminals. The relationship between the three businesses is very vertical, and the user groups and product forms are so different that it is difficult to unify them with one existing name. Internal discussions came and went, and finally Alimama came to mind, and the brand was relaunched.
Wu Xuejun has led the Alimama technical team for a short time. In 2013, Ali put forward the "all in wireless" strategy and began to draw manpower from various departments of the group to develop hand-to-hand Tao, and Wu Xuejun was among them. Welcome to add the author's WeChat LW_PLUS to learn more about the story behind Alibaba's "all in wireless".
According to the instructions of his superiors, Wu Xuejun could select 5-6 elite generals from his subordinates to follow him to Shoutao. So he immediately found Yuan Quan, who was in charge of the recommendation advertising algorithm, who not only readily agreed, but also called his right-hand man Yan Qiang.
Subsequently, Wu Xuejun successively persuaded Zhao Binqiang (Le Tian), Zhou Liang, and Wang Zhe to form a six-person team from Beijing to Hangzhou to help with the construction of hand-to-hand Tao.
In terms of the selection of team personnel, Wu Xuejun has two criteria: one is experienced, and the other is capable and potential.
In the six-person squad, Yuan Quan and Zhou Liang are both experienced veterans. Among them, Yuan Quan used to work in IBM Research Institute and was one of the first people in China to make recommendation algorithms. During his time at IBM, he and Xiang Liang, who was an intern at IBM at the time and later became the head of the byte recommendation system, published KDD papers together.
Zhao Binqiang is also very senior, he joined Alibaba in 2008, and has been in B2B and Alibaba Cloud and other businesses, and later came to Alimama, which is relatively balanced in terms of engineering and algorithms.
Although Yan Qiang is young, he is also very sophisticated. In 2008, Wendong Gu, a technology leader in the field of recommendation, and Xiang Liang, who is still studying for a PhD, founded ResysChina, a professional community for the field of recommendation, and often holds small gatherings. Yan Qiang, who was still a graduate student at the Chinese Academy of Sciences at the time, was a frequent visitor to community gatherings.
Wang Zhe belongs to the category with strong hands-on ability, which just complements others.
As for Gai Kun, who is known as the genius of Ali's algorithm, although he is also a rare young talent, his research direction at that time was more theoretical, and it was a little far away from the recommendation algorithm, so Wu Xuejun could only give up.
In November 2013, Wu Xuejun led a team of 6 people from Beijing to Hangzhou and began a difficult technical battle. Alibaba's recommendation team has also been divided into two, one part continues to be responsible for advertising recommendations at Alimama, and the other part is based on organic traffic on hand-to-hand Tao.
During that time, the team worked six days a week, from 9 a.m. to 9 p.m., mostly working except for eating and sleeping. With the rest of the day, everyone will sit together in the dormitory to play DOTA and release the pressure. It was at that time that the buzzword 996 in the Internet industry began.
After Wu Xuejun's team came to Handtao, the first personalized product they made was "good goods".
At the beginning of 2014, Daniel Zhang took over the banner of Hand Tao from Wu Yongming. This heavy burden forced him to think about what to do every day. He was even obsessed with printing out the pages of the first three screens of the hand-to-hand shop, putting them on his desk, and figuring out which module should be placed in that position every day. Later, Jiang Fan gave him an idea, saying that since he didn't know what kind of page could be loved by the most users, he would simply personalize it, so that each user could see a different page.
The first renovation project of hand-shopping personalization is "there are good goods".
This product was not valued by the top management at first,It was in a very marginal position in the hand Tao page,Later, after the personalized recommendation algorithm,All indicators are very good,It was moved to the middle of the first screen of hand Tao。
The success of "Good Goods" strengthened Daniel Zhang and Jiang Fan's determination and confidence to make personalized recommendations, and they decided to extend the recommendation algorithm to the second and third screens.
At that time, Facebook's feeds had already emerged, so "how to do e-commerce feeds streams" began to be put on the agenda of mobile Taobao.
After many internal discussions, it was decided to make two columns of product feeds - which later became "Guess You Like It". Jiang Fan is responsible for promoting from the product and business level, and Yuan Quan is responsible for the back-end algorithms.
"Guess You Like" started by simply reusing the "Good Goods" algorithm and only displaying 10 products. But later the team found that users like to brush "guess what you like" when they are bored on the bus and subway, and the 10 products were quickly brushed, so they decided to increase it to 20 or 50, but it turned out that it was still not enough, so they simply set it to "infinite display".
At the beginning of 2014, "Guess You Like" had only one or two million users when it was first launched, and by 2017, it had exceeded 100 million daily users, becoming the largest recommended product among all e-commerce.
The technology and experience accumulated by "Guess You Like" were later brought to various e-commerce platforms with the turnover of the team.
The battle of "all in wireless" won a ticket to the mobile Internet for Ali. After the end of the war, the recommendation algorithm team led by Wu Xuejun was merged into the search division at the same level as Alimama, and reported to Zhang Qin (Yang Guo), the head of the department. Recommendation algorithms and search teams are gradually converging.
Soon, Gu Xuemei parachuted into Alibaba from Google China Research and took over the search division. Around the same time, Wu Xuejun resigned from Ali. Alibaba's search and push team has since entered a new era.
Jiang Long laid a solid foundation, and the recommendation algorithm began to rise
Turning the pointer back to the end of 2013, after Wu Xuejun led Yuan Quan and others to Hangzhou, Alimama's technical team was divided and conquered by Xue Guirong (Da Shu) and Tang Yong (17), Tang Yong was responsible for the search, and Xue Guirong was responsible for the rest.
Tang Yong is an old Ali, who was born in system architecture, and was responsible for building Alibaba's first-generation advertising system architecture, and is very familiar with Wu Yongming, Liu Zhenfei and other executives. Xue Guirong is the only technology master who has won two Microsoft Scholar Scholarships, joined Alibaba in 2009, did a search in Alibaba Cloud, and was later transferred to Alimama, leading the team to make one of the most important marketing tools of Alimama - Dharma Pan (DMP).
At that time, Alimama's algorithm team was divided into four lines – Search, Recommendation, Targeting, and Tanx. Among them, Yuan Quan is responsible for recommendation, Gong Bihong (Skylark) is responsible for searching, and Jiang Long is responsible for orientation. The Tanx team is mainly responsible for brand advertising, and the team is based in Hangzhou, which does not use many advanced technologies. After Yuan Quan followed Wu Xuejun to support wireless, the recommendation advertising algorithm that he was originally responsible for was handed over to Jiang Long. In this way, Jiang Long, who holds the two lines of recommendation and targeted advertising, has become the essence of the core of the Alimama algorithm team.
Jiang Long
Jiang Long and Wu Xuejun are old acquaintances. As early as 2009, when Wu Xuejun was still in Baidu, he found Jiang Long, who worked at Microsoft Research Asia, through headhunting, and threw an olive branch to him. But Jiang Long felt that Baidu was not suitable for him, so he politely declined.
Two years later, Wu Xuejun, who had already joined Alibaba, sent another invitation to Jiang Long to join. The two met in the Taobao office of the Taikang Building on Beijing's East Third Ring Road. At that time, Jiang Long, who was at Microsoft Research Asia, had already recognized the shortcomings of the research institute model - it was too far from industrialization.
In the traditional software era, the software was engraved in a disc and sold to the user, and if the software had a bug, the user could not return the disc. Therefore, traditional software vendors, represented by Microsoft, have a very strict product development process - at least one year of planning, two years of development, and two years of testing. It takes at least five years for a technology to mature and reach users.
Therefore, although many of Microsoft's technologies were very advanced at the time, they often fell behind the times when they were productized.
In contrast, Google's Internet model of "small steps, fast iterations". This efficient industrialization method caused a deep shock and impact on Jiang Long, and gave him the idea of going to an Internet company.
In fact, before this, Microsoft Research Asia has lost many talents to Internet giants and achieved great success, such as Wang Jian of Alibaba Cloud. This further strengthened Jiang Long's idea of leaving.
However, Jiang Long did not follow in Wang Jian's footsteps to join Alibaba Cloud, but had other plans.
Jiang Long wants to do artificial intelligence, and the first thing to do is data. Search and advertising are both data-rich scenarios, but there is a big difference between them: at that time, search was not profitable, it was just a product that improved the user experience, while ad optimization directly led to a revenue increase.
Jiang Long believes that only if the business can make money for the company, the company will continue to invest in it. Therefore, when Wu Xuejun invited him to join Alimama, he chose to join with little hesitation. He knew that this must be a business that Ali would continue to invest in.
I have to say that Jiang Long's eyes are indeed vicious. With the blessing of the algorithm, Alimama's revenue soon ushered in explosive growth, once accounting for more than eighty percent of Alibaba's overall revenue.
It is said that after 2013, because the revenue growth exceeded the budget too much, at the end of each year, Alimama had to control the growth of advertising revenue by reducing advertising space and shrinking advertising pictures.
With a steady stream of cash inflows, Alimama naturally invests heavily in technology.
In the early days, Alimama's advertising model was based on established rules, and product managers needed to plan various rules such as sorting in advance, and then implement them through algorithms. This is not only inefficient, but also a compromise in the design of the rules. In order to solve this problem, Jiang Long joined and led the team to build a truly data-driven, large-scale machine learning system.
In order to build this system, Jiang Longxun applied for a budget of several hundred million yuan to build an Alibaba's most advanced computing power cluster. At that time, a server cost hundreds of thousands, and the group directly purchased 500 units with a wave of its hand.
Alibaba's executives don't really understand the details of the algorithm, but Jiang Long told them, "The hundreds of millions you invest can bring billions of revenue growth a year." Executives knew right away what to do.
On the other hand, the search team that Yuan Quan went to was not so lucky. Because the monetization path of search is relatively long, they can't say what kind of benefits technology investment can bring, so it is difficult to get more computing power. For a while, I could only borrow computing power from Alimama. Welcome to add the author's WeChat LW_PLUS to learn more about the interesting stories behind it.
The large-scale machine learning system built by Jiang Long has helped Alimama a lot, but his greatest contribution is that he has completed the formulation and training of Alibaba's algorithm talent standards with the algorithm committee.
There were not many people who knew algorithms in the early days of Ali, and Wu Xuejun and Jiang Long were the earliest batch of algorithm talents. The two, together with Xu Yinghui (Renji) and others, established the first algorithm committee of Alibaba Group.
The algorithm committee has mainly done several things: First, it has formulated Alibaba's algorithm talent standards, such as how to grade talents after they are introduced, and what are the promotion standards for P7/P8/P9. The second is to build Alibaba's algorithm system, sort out the algorithm technology directions involved in internal business, and what problems are involved in each direction, which is equivalent to building a knowledge sharing system internally. The third is to cultivate a group of outstanding algorithm talents for Ali, and the members of the committee have personally written teaching materials and held classes, and now many of Ali's algorithm backbones have benefited from it.
These laid a solid foundation for Alimama to climb to higher heights later.
In 2014, Alimama had entered the fast lane of development, but Jiang Long decided to leave on the eve of promotion.
In his opinion, although Alimama was in full swing at that time, and its revenue could maintain rapid growth for at least five years, the iteration of advertising system technology itself was very slow, and he should find a broader world to display his talents.
After Jiang Long left, he handed over the scepter of the recommendation algorithm team to Gai Kun.
Jiang Long felt that Gai Kun was very similar to himself, and he was the same kind of person in his bones. Both of them have a very advanced technical vision and can see the development trend of technology one step ahead of others. Moreover, Gai Kun has a relatively gentle personality and a weak desire to control, and can give his subordinates a high degree of freedom while providing enough support. Colleagues commented that Gai Kun, although he is not the kind of enthusiastic and especially good at getting things done, but his emotional intelligence is definitely not low, and he can very keenly capture the other party's emotions and react. These personalities and traits give him the opportunity to be a great team leader.
Of course, more importantly, Gai Kun's achievements in Alimama are enough to convince the public.
Gai Kun took the scepter, and Ali Mama stepped into the pinnacle
The process of Gai Kun joining Ali is quite dramatic.
Gai Kun is the top student in the college entrance examination in Gansu Province. Prior to joining Alibaba, he studied for a Ph.D. at Tsinghua University, focusing on machine learning and computer vision. When he was still in school, Gai Kun had already made small achievements in the academic world, published many academic papers in top international journals and conferences, and won the Excellent Doctoral Dissertation Award of the Chinese Association of Artificial Intelligence.
In 2011, Ali held a recruitment seminar at Tsinghua University. At that time, Gai Kun was thinking about research institutes such as MSRA and IBM CRL, and he was not interested in joining the company, so he did not participate.
Later, Gai Kun happened to see a recruitment post posted by Wu Xuejun on the Shuimu Tsinghua forum, which said that Ali was working on a large-scale machine learning thing. This piqued his interest, so he replied with a private message.
After replying to the private message, Gai Kun played the game as usual and played until two or three o'clock in the morning before going to sleep. As a result, I was woken up by Alibaba's HR at eight or nine o'clock the next morning to go for an interview.
After several rounds of interviews, Gai Kun found that Ali had a strong determination in machine learning, and the business demand for related technologies was particularly strong, so he happily chose to join.
Of course, the conditions given by Ali are also very tempting, and he was directly given P7, while most fresh graduates are P6. On the one hand, this is because Ali values Gai Kun's potential, and on the other hand, because his doctoral project is very similar to Alibaba's advertising algorithm, and the results can be directly transferred to the application. And Gai Kun also lived up to expectations, a year later to P8, another two years to P9, another three years to P10, is one of the fastest people to upgrade to P10 Ali.
Gai Kun
In 2011, shortly after joining Alibaba, Gai Kun broke through the mainstream large-scale linear model and creatively proposed the MLR (mixed logistic regression) algorithm, which made him known as an "algorithm genius" in one fell swoop.
The MLR algorithm innovatively proposes and realizes the nonlinear relationship between features directly learned in the original space, and automatically discovers generalizable patterns based on data, which greatly improves the efficiency and accuracy compared with manual work.
Since 2013, MLR algorithms have been applied and tried on a large scale in the main scenarios of Alimama and multiple BU of Alimama Group (including Alimama precision targeted advertising, Taobao Ke, WhatsMiner commercial advertising, Taobao main search, etc.), especially in the precise targeted advertising scenario of Alimama, the algorithm model innovation has brought a major breakthrough in business, and the CTR and RPM in the main scenarios have been improved by more than 20%.
Unfortunately, MLR algorithms have not been adopted on a large scale in the industry. On the one hand, this is because Ali did not open source the MLR algorithm in the form of a paper until 2017, so the details of the implementation of the MLR model were not clear to the outside world for a long time. On the other hand, the solution of traditional LR+ feature engineering has deeply influenced the way of thinking and organizational structure of many technical teams.
In 2017, the Alimama Precision Directional Retrieval and Basic Algorithm Team led by Gai Kun proposed another important recommender system model, Deep Interest Network (DIN), which applied deep learning to Alimama for the first time. The model proposes that users' interests are diverse, and uses deep learning to make full use of/mine the information in users' historical behavior data to improve the performance of CTR prediction.
With the successive introduction of algorithm models such as MLR and DIN, Alimama has been at the forefront of the industry in terms of revenue scale and technical influence in those years.
A side example is that around 2015, Byte poached Ali's algorithm team on a large scale, but the success rate was not high. The reason is that at that time, Ali was very competitive in terms of team atmosphere, paper output speed, work fulfillment and remuneration.
According to the recollection of Alimama members at the time, "even if you are a P5 intern, you can publish papers soon after joining."
But for Gai Kun, despite becoming a notorious tech genius with a few key technological breakthroughs, he also has his own troubles – a lack of control over the business.
Any technological breakthrough needs to fall into a specific business scenario to be valuable, otherwise it is just a bunch of code composed of 0s and 1s.
Alimama's most profitable business has always been the through train, but the through train is not in the hands of Gai Kun, he is only responsible for the accurate display of advertising. Moreover, the ecology on which accurate display advertising depends, such as guess what you like, is also in the hands of the recommendation team of the search division, and Gai Kun does not have the absolute right to speak. Even the search division, in turn, is eyeing the business in Gai Kun's hands.
At the beginning of 2017, Wang Xiaobo (Uncle Yong) of Alibaba's search division began to lead the team to do information flow (Wang Xiaobo now serves as vice president of technology at Xiaohongshu). With the information flow, Gu Xuemei, then head of the search division, naturally thought of opening advertising space to do business closed-loop. Although Alimama has always been in charge of this part of the work, Gu Xuemei believes that the search department can do the same, or even better.
Because of this incident, the contradiction between the search division and Alimama has intensified. After the conflict escalated to the top, it was decided internally to let the two teams have a positive PK, so that both sides could do a part of the information flow and advertising recommendations at the same time, and see who had the better effect through bucket testing. Whoever wins the business goes to whomever it is.
Alimama sent Gai Kun here, and the search division was led by Wang Xiaobo, and a PK of needles against Mai Mang began.
Dramatically, there is an employee in the search department who has put some thought into product design in order to increase his chances of winning. She designed a small assistant button in the lower right corner of the feed page, so that users can jump to their favorite position with a single click if they don't like the previous content. This button can be controlled by an algorithm to appear frequently, and this employee moved his hands and feet in this link, so that the frequency of the assistant button appeared in the test barrel where Gai Kun was located was lower.
Unexpectedly, this cheating behavior was caught by Ali's mother, and she was directly stabbed by the Ministry of Anti-Corruption. During the investigation by the Ministry of Integrity, Gu Xuemei, who is the head of the search department, argued that the actual impact of the cheating on the results was very small (the measured result was only 3‰), and it was not worth fighting. But in any case, cheating in PK, the search division always loses first.
In the end, the high-level discussed it, and felt that since Gu Xuemei liked advertising so much, and even quarreled with Alimama for this, it was better to let her go to Alimama.
In this way, Gu Xuemei joined Alimama with the recommendation algorithm team in the search division and became the head of technology. The trade-off is a loss of control over the search team.
With Gu Xuemei's arrival at Alimama, the recommendation algorithm team, which was split into two at the time of "All in Wireless", has moved towards integration again. Unfortunately, Gu Xuemei only stayed at Alimama for a short time before leaving.
She and Zhu Shunyan, the president of Alimama at the time, had great differences in the way of employing people, and the relationship has not been harmonious. Gu Xuemei is a technical person who likes to be systematic, theoretical and organized, so Gai Kun is highly regarded by her. Zhu Shunyan is a marketing background, once responsible for the marketing and commercialization system of UC browser, he advocates customer first, and pays more attention to Hu Yunhua (Wu Hook) of Taobao live broadcast car.
The difference in ideas makes the two often have antagonistic emotions when they meet. Gu Xuemei was very unhappy in Alimama because of this, and it happened that at that time, rookie CTO Wang Wenbin (Fei Qing) retired, and the rookie needed a new technical leader, so Gu Xuemei turned around and went to rookie as CTO. Welcome to add the author's WeChat LW_PLUS to learn more details about the story behind it.
Later, at a rookie conference, someone asked Gu Xuemei why she left Alimama, and she bluntly said, "I can't figure out the person in charge of search advertising."
Zhu Shunyan also stepped down as the president of Alimama soon after and became the president of Alimama UC Business Group. Alimama began to change coaches frequently, and successively experienced the era of Dong Benhong (Zhang Wuji), Zhang Yifen (Zhao Min) and Jiang Fan's management.
postscript
From 2010 to 2017, Alimama was the fastest growing years, and Alimama was an important engine to provide surging power.
During this period of time, Ali's performance was thriving, looking at the stars and the sea, providing a stage for a large number of talented and ambitious technical talents like Wu Xuejun, Jiang Long and Gai Kun to display their talents.
The stars complement each other, creating the golden age of Alimama, and the broad stage of Alimama also makes their light visible to more people.
Today, many of these people have left Ali and brought the vision and experience they have accumulated here to more companies and fields, scattering the stars. This is the inheritance and continuation of the spirit of the Ali people.
And Alimama is still the most important money bag of Alibaba Group and the core of the business in charge of traffic distribution. At present, Alibaba is in the midst of drastic changes, and how to better allocate traffic to achieve optimal efficiency and user experience is the core and key proposition of this change.
Last year, Alibaba put forward two strategies: "user-first and AI-driven". As the source of Ali's AI culture and the eye of the storm of current change, Alimama carries the expectations of countless people.
Everyone is looking forward to the fact that this business, which has brought countless glories to Ali, can help Ali achieve the return of the king and write new glories in the era of AI e-commerce.
Regarding the story of Ali's AI development, Leifeng.com will continue to launch a series of articles on Ali's AI-driven 20 years, "Those who have searchers win the world, the prequel of the Damo Academy", "Whose middle platform is the middle platform, whose Damo Academy is the Damo Academy", "In the post-Damo Academy era, the AI-driven strategy of Ali Group", interested readers should contact the author to communicate (WeChat LW_PLUS). Leifeng.com