Doubao Big Model Pre-training Product Operations

1. Collaborate with the algorithm team to jointly formulate pre-training data processing strategies based on the current status of the big model 2. Confirm the ideal state of the data based on the strategy, and filter and process the pre-training data through PE and other means 3. For the trained model, sort out problems and evaluate the effect 4. Optimize the training process and resource allocation to improve training efficiency.

1. Bachelor degree or above, more than 2 years of experience in Internet products or operations, English can be used as a working language, PE experience is preferred 2. Have a strong sense of content, computer, education, law, psychology, physics, chemistry, biology, screenwriting and other majors are preferred 3. Interested in the field of large language models, willing to devote themselves to the active exploration of large language models/AIGC and other fields 4. Have a strong sense of responsibility and stress resistance, good communication and teamwork skills, and be able to complete work tasks on time and with high quality 5. Bonus points: experience in the field of large models and understanding of the training logic of the model.