In 2019, did everyone notice a social phenomenon where competition for AI positions became increasingly fierce?However, at the same time, companies say that there is a severe shortage of AI talent?This seems a bit contradictory, but upon closer examination, the main reason is that many people want to enter the high-paying AI industry, but the vast majority do not meet the requirements for these positions. A few years ago, if you were proficient in TensorFlow and had a grasp of basic AI algorithms, you could easily find a high-paying job, but now it is different; the requirements for AI positions are getting higher, demanding deeper knowledge.
01
Three Major AI Professional Directions
As major AI companies lay out their artificial intelligence strategies, today’s recommendation systems for platforms like Toutiao and Taobao, Google’s search engine, voice assistants like Siri and Tmall Genie, Baidu’s self-driving cars, and Meitu’s AI beauty technology all apply natural language processing, computer vision, machine learning, and other AI technologies.The course will focus on these three directions, starting from basic to advanced with enterprise-level project practice, aiming to secure an AI job offer in 4 months.
Intensive Learning Schedule
<<<Swipe to view the three major professional course outlines
<<<Swipe to view the three major professional course outlines
Please long press to scan the QR code below
Add the Greedy AcademyAI Career Planning Consultant
Consult for course details
For your learning and career planning
Enterprise-Level Project Practice
Unlike ordinary demo explanations,each project needs to be completed independently or in collaboration with team members, averaging about 1 month to complete.This increases practical experience while enhancing the value of projects on your resume.
Project 1: Knowledge Graph-Based Q&A System in Healthcare
[Project Overview]
The healthcare + AI field has always been a hot topic in the industry. In this project, everyone will have the opportunity to complete a Q&A system based on a knowledge graph. From building the knowledge graph to the Q&A system module, and finally to deployment, you will experience the complete process.
[Technical Points Involved]
Web scraping, text data cleaning, preprocessing, text mining, text classification, knowledge graph construction and querying, sequence-to-sequence models, attention mechanisms, model deployment.
Project 2: Generate Text from Structured Data
[Project Overview]
In the industrial sector, the application of text generation is everywhere, such as automatically generating weather forecasts, sports news, and student ability reports, etc. In this project, you will have the opportunity to generate a piece of text based on given structured data.
[Technical Points Involved]
Pre-trained models, seq2seq, attention mechanisms, copy mechanisms, loss function optimization, text quality assessment (rouge, bleu, etc.)
Project 3: Kaggle Competition (Text Matching)
[Project Overview]
Quora can be considered the English version of Zhihu, and matching users with similar questions is a challenge faced by the platform. Customer service robots and other applications also use text matching technology.
[Technical Points Involved]
Writing deep learning modules (LSTM, CNN, FC, Dropout, etc.), feature engineering, model tuning, model integration, text matching models.
Project 4: Object Detection and Segmentation Based on Deep Learning
[Project Overview]
Object detection is one of the most researched and widely applied branches in the field of computer vision. Object detection is widely used in face detection, autonomous driving, medical image processing, and many other fields. The segmentation technology aims to provide a more detailed description of the contours of detected objects, which is a more precise identification method than detection. In this project, we will perform both object detection and segmentation in the same algorithm, optimizing the model for faster speed and higher accuracy.
[Technical Points Involved]
Image datasets, pre-trained models, image annotation, bounding box regression, loss function selection, model quality evaluation.
Project 5: Human Keypoint Detection
[Project Overview]
Human keypoint detection can be used to estimate the posture of the body and hands. Estimation of keypoints can be applied in human motion analysis, behavior analysis, drawing assistance, game character motion capture, virtual reality, and other fields. In this project, we will build a human keypoint detection system and output detection results in real-time.
[Technical Points Involved]
Image preprocessing, Hourglass Networks, point regression, point classification, relation matching, Hungarian algorithm.
05
Course Development Instructors
▲ Some Course Development Instructor Profiles
06
Course Features and Teaching System
The course adopts a live teaching method, with 4 teaching arrangements per week, including 1 professional course, 1 public course, 1 project case explanation (explaining a practical case or technical extension), and 1 paper reading session (guiding and explaining a must-read paper).The teaching model refers to the teaching training system of top American universities.The following is a weekly course schedule for reference.
No matter how many obstacles you encounter during the learning process, you can resolve them through the following 4 methods:
-
Directly ask the instructor online;
-
Or record it in a shared document and wait for the fixed time of live Q&A
-
Full-time teaching assistants in the learning community, available for questions 24/7
-
Common questions will be explained in the Review Session
Note:Each Q&A session will be recorded by the head teacher for students to review in real-time.
③ Publish Technical Articles:
By publishing relevant technical articles on Zhihu to test your achievements, it is also a way of intellectual collision. The instructor will write a detailed comment on each article published.What if you accidentally become a big influencer?
Although writing articles is extremely painful, and the learning group often wails at midnight, looking at the results of the articles written while pulling hair is still very gratifying!Seeing the number of likes you receive, everyone silently thanks the ruthless instructor!
This sense of achievement makes everyone keep writing one after another!Everyone immediately transforms into a Zhihu expert.
In addition to articles, the fundamental aspect of an algorithm engineer’s life—project code.For each assignment, the instructor and the assistant team will provide detailed corrections and feedback.And they will force you to keep optimizing!
Student Employment Situation
▲ Some Student Employment Situations
08
Scholarship and Tuition Information
The content of the Pass Class and Worry-Free Class is the same, but the Worry-Free Class will sign a guarantee agreement on a third-party notarization platform.Guaranteed placement.
For students who meet the following conditions, we provide a scholarship reward of 50% of the tuition:
1. A master’s or doctoral degree from a university ranked in the top 100 by QS or US News, or currently pursuing a master’s or doctoral degree;
2. Have published at least one paper as the first author, or two or more papers as a non-first author in CCF C-level conferences (including C-level);
3. At least one project in your own Github with more than 300 stars;
4. Achieved first prize or higher in the Asia region or international competitions such as ACM, MCM;
5. Achieved top 5% in competitions like Kaggle and Tianchi;
Students enrolling in the Worry-Free Class will sign an agreement at the start of enrollment,promising:If you do not receive an AI job offer or your initial pre-tax salary is less than 220,000/year (160,000/year for undergraduates), the full tuition will be refunded.
Please long press to scan the QR code below
Add the Greedy AcademyAI Career Planning Consultant
Consult for course details
For your learning and career planning
