AI Intern
Job Description
- Responsible for large model training in image and video domains.
- Design and implement efficient distributed training architectures.
- Evaluate and validate visual models using metrics like mAP, IoU, accuracy.
- Track and apply the latest research in computer vision and deep learning.
- Optimize hardware resources for efficient training.
- Collaborate with business teams to deploy visual intelligence solutions.
- Master's student in AI, Machine Learning/Deep Learning, Computer Science, or Applied Mathematics.
- Solid deep learning theoretical foundation and practical model training experience.
- Proficiency in mainstream AI applications like OpenAI and DeepSeek.
- Proficiency in C/C++/Python in a Linux environment.
- Mastery of deep learning frameworks such as TensorFlow and PyTorch.
- Good work documentation habits.
- Ability to work independently, good teamwork spirit, and continuous model tuning skills.
- Available for an internship of 3 months or more, with a minimum of 4 days per week.
Job Description
1. Responsible for large model training in the image and video domains, including image/video dataset preprocessing, annotation quality control, model architecture design, and training process optimization, ensuring high accuracy and reliability of models in visual tasks.
2. For tasks such as image classification, object detection, semantic segmentation, and video behavior analysis, design and implement efficient distributed training architectures, fully utilizing company computing resources to accelerate model training and process large-scale image and video datasets.
3. Conduct multi-dimensional evaluation and validation of trained visual models, measuring model performance through visual task metrics such as mAP, IoU, and accuracy, and promptly adjusting training strategies combined with visual analysis to optimize model effects in image/video scenarios.
4. Track the latest research achievements in computer vision and deep learning, focusing on technical directions such as image generation, video understanding, and multimodal fusion, applying new technologies to model training, and continuously improving the performance and innovation capabilities of large visual models.
5. For the storage, preprocessing, and training characteristics of image/video data, reasonably allocate hardware resources such as GPU and CPU, optimize training costs, and ensure the efficiency and economy of large-scale visual data training processes.
6. Work closely with business teams to promote technology implementation for actual image/video business scenarios (e.g., intelligent monitoring, content review, visual retrieval), improving the department's work efficiency and service quality in the field of visual intelligence.
Job Requirements
1. Currently pursuing a Master's degree in Artificial Intelligence, Machine Learning/Deep Learning, Computer Science, Applied Mathematics, or related fields.
2. Possess a solid theoretical foundation in deep learning and certain practical experience in model training, proficient in various mainstream AI applications, such as OpenAI and DeepSeek series tools.
3. Proficient in C/C++/Python programming languages in a Linux environment, mastering deep learning frameworks such as TensorFlow and PyTorch, capable of flexibly using frameworks for model training and optimization.
4. Have good work documentation habits, timely writing and updating work processes and technical documents as required.
5. Possess independent problem-solving abilities, good teamwork spirit, capable of continuous model tuning, and familiar with methods for model tuning.
6. Available for an internship of 3 months or more, with a minimum of 4 days per week, and able to start as soon as possible is preferred.
