Job Summary
The primary focus of this position is to cultivate and expand diverse datasets comprising text, code, images, and videos intended for the pretraining of multimodal AI models. The role demands a strategic approach to devising filtering techniques, implementing synthetic supervision, and executing controlled experiments aimed at enhancing the quality of the models and the efficiency of the training process.
Responsibilities
- Develop and refine text, code, image, and video datasets for multimodal AI model pretraining.
- Design and implement filtering strategies to enhance dataset quality.
- Create and manage synthetic supervision mechanisms.
- Conduct controlled experiments to optimize model training efficiency and performance.
- Collaborate closely with teams to ensure data strategies align with overall project goals.
Qualifications
- Demonstrated experience in creating training data for language or vision-language models.
- Strong proficiency in Python programming.
- Experience with distributed processing frameworks.
- A proven track record of enhancing model performance through comprehensive data research methodologies.
Hark provides equal employment opportunities to all employees and applicants for employment and prohibits discrimination and harassment of any type without regard to race, color, religion, age, sex, national origin, disability status, genetics, protected veteran status, sexual orientation, gender identity or expression, or any other characteristic protected by federal, state or local laws.
