Interview notes:

Su Yu: Luo Jingyuan, an expert in labeling training, was invited for this interview with Hokkestar. The rematch just came to the stage of labeling. Could Mr. Luo introduce your experience in labeling work?

Luo Jingyuan: Hello, Mr. Su, my experience can be traced back to the time when I had not graduated from university. At the beginning, I was doing customer service work and doing job rotations. During this process, I accumulated a lot of experience and learned about computers and artificial intelligence. I saw some hope in the development of intelligence, so I had the opportunity to participate in the development and operation of an internal intelligent robot product. At that time, I didn’t have a clear concept of artificial intelligence. I just thought that robots can solve customer problems very well. Later, I found that robots The growth of the need to collect a lot of corpus, and some cleaning, labeling, analysis of these corpus…

Su Yu : How do you understand the profession of artificial intelligence trainer from the aspect of data labeling? What are the specific tasks of artificial intelligence trainer positions?

Luo Jingyuan: Speaking of data labeling, we can first look at the entire artificial intelligence product development, which can be divided into three links: 1. Computing power; 2. Algorithms; 3. Data. In fact, 70%-80% of the time in the research and development of artificial intelligence is spent in data processing. According to this classification, the personnel we need include data labelers and algorithm testers…

Su Yu: At present, China recognizes artificial intelligence trainers very much. Everyone is engaged in this industry. As the founder of the company, what do you think about the employment prospects of artificial intelligence trainers in this industry?

Luo Jingyuan: It is no exaggeration to say that I think artificial intelligence trainer is a career that can be used for a lifetime. I think its employment prospects are particularly good. I have made preliminary calculations, and I think this industry needs at least 10 million talents. People, these people can participate in the process of data labeling and cleaning, and other related work also requires the addition of talents in the industry of artificial intelligence trainers. So at present, I think that the profession of artificial intelligence trainer is very easy to find employment. Like the new infrastructure proposed by the country, artificial intelligence is one of the seven major themes…

Su Yu: Thank you very much, Mr. Luo. I heard that your company is also doing some overseas business. Can you briefly share with us?

Luo Jingyuan: Well, Mr. Su, in fact, we have two paths. One is to help domestic enterprises go overseas, and the other is to meet the real needs of some foreign industries. We need to integrate the way foreigners adapt to our industry, use a model that is more suitable for local humanistic development to simulate application scenarios, and make related businesses land. This requires a lot of image and video collection. Through some cameras, some Video recording to understand the local environment, and then through voice collection and spoken language collection to make smart products more local characteristics…

Su Yu: Okay, soon our Future Star Contest will enter the semi-finals. This stage is mainly to compare the accuracy of labeling. As an expert in this field of labeling, can you share with other partners if you participated in this stage of the competition? What are the elements of the management of the labeling process?

Luo Jingyuan: Good Mr. Su, the processing of labeling is the main task of our data work. The first thing to avoid is to process data without understanding the data rules, which will cause the data to be repaired; so the first step we need to understand Processing requirements and mastering processing rules; in the second part, we need to collect a lot of samples, and these samples should cover various scenarios of the data we have made, so that under the condition of controlling the rules, the data produced can effectively reduce repairs the time required…

Su Yu: Alright, thank you for watching this episode of the Hospitable Star interview. I also wish all contestants the best results in the competition.

Follower: Qian Yi

Written by: Qian Yi

Planning: Su Yu

Time: June 2022

Location: Online

Team introduction: Captain: Luo Jingyuan, team members: Wang Le, Zhang Lang, Jiang Jing Chengdu Huizhong Tianzhi Technology was established in 2017. It is a national high-tech enterprise and the only artificial intelligence data service unit participating in the “National Occupational Skills Standards for Artificial Intelligence Trainers” business.