Data Scientist

Data Engineer - Video Generation

About Us

We are a cutting-edge AI startup specializing in next-generation video generation technology based in Hong Kong. Our mission is to push the boundaries of what's possible in AI-driven video generation through innovation of foundation model. As a growing startup, we offer a dynamic environment where your research can have immediate impact on technology development.

Position Overview

We are seeking a skilled Data Scientist to design, develop, and optimize foundational AI models and multi-modal data processing systems. The ideal candidate will have strong experience in researching and building in handling large-scale multi-modal datasets (videos, images, texts) for advanced AI applications.

Key Responsibilities

Design and implement AI models for aesthetic evaluation and video quality detection, and iterate models based on business scenarios and evaluation results

Build and optimize multi-modal data processing pipelines for model training, including data cleaning, annotation, augmentation, and feature engineering for videos, images, and texts

Collaborate with engineering teams to deploy models into production environments, and provide technical support for model iteration and performance tuning

Conduct technical research on cutting-edge AI technologies (especially multi-modal large models and video generation models) and apply them to practical business scenarios

Work closely with product and business teams to translate business needs into data solutions and drive the implementation of AI technology applications

Required Qualifications

Experience in developing multi-modal understanding models (e.g., text-image, text-video cross-modal models) or video generation models

Practical experience in building and optimizing aesthetic evaluation models or video quality detection systems

Experience in processing and utilizing large model training data (data curation, deduplication, quality control, etc.)

Familiarity with distributed training technologies for large models and experience in improving model training efficiency

Conduct in-depth data analysis, including model performance data, user feedback data and business operation data, to provide data-driven guidance for model iterative optimization and commercialization strategy formulation

Preferred Qualifications

Experience with video processing

Publications in top AI conferences (NeurIPS, ICML, CVPR, ICCV, ECCV, ACL, etc.) or industry awards related to AI model development

Bilingual proficiency (English/Chinese)

Master's or Doctor's degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, or related field

Technical Skills

Strong programming skills in Python, and proficiency in deep learning frameworks (PyTorch, TensorFlow, MindSpore)

Proven track record in developing foundational models for computer vision (image/video understanding) or natural language processing (text understanding)

Model Architectures: CNN, Transformer (Vision Transformer, BERT, GPT, etc.), multi-modal fusion architectures

Experience in handling large-scale multi-modal datasets (videos, images, texts) and building data processing workflows for model training

Model Tuning: Fine-tuning, few-shot learning, transfer learning, quantization, pruning

Strong understanding of machine learning principles, deep learning architectures, and model optimization methods

What We Offer

Opportunity to build critical infrastructure for cutting-edge AI technology

Competitive salary and equity package

Flexible work arrangements

Professional development opportunities

Modern tech stack and tools

Collaborative and innovative work environment

Health and wellness benefits

Expected Impact

Shape the core AI technology foundation of the company's products through foundational model research

Influence key technical decisions on multi-modal AI architecture and model roadmap

Build and mentor a world-class AI technology team and drive technical innovation

To Apply:

Please submit:

Detailed CV with publications and major projects

Brief description of the most complex data pipeline you've built

Links to any open-source contributions or technical blogs

To apply or learn more about this position, please contact [HIDDEN TEXT].