Full timeHefei,China
Multimodal Algorithm Research Enginee
Job Objective
Research and develop multimodal generative models to achieve breakthroughs in human–robot interaction experience and computational efficiency.
Job Responsibilities
- Research and advance the application of multimodal models in face-to-face interactive robots; explore and build innovative algorithm models and technical solutions.
- Deeply participate in data construction, and design, develop and optimize multimodal generative models.
- Keep track of cutting-edge advances in multimodal AI technologies, and conduct technical research and summary.
Job Requirements
- Master’s degree or above in Artificial Intelligence, Big Data, Computer Science or related fields. First-author publications in CCF-A journals/conferences or equivalent venues on topics including speech synthesis, image/video generation, VLM, audio large models, etc.
- Proficient in architecture design and training optimization of multimodal generative models (e.g., VLM, audio large models, video generation models); solid command of mathematical principles and training strategies for cutting-edge frameworks such as Transformers and diffusion models.
- Familiar with multimodal dataset construction workflows and quality standards; capable of developing scripting tools for dataset construction and data cleaning.
- Master systematic model evaluation methodologies and inference optimization techniques.
- Experienced in complete research projects; able to clearly elaborate scientific problems, research approaches, technical routes and core conclusions in a structured manner.
- Passionate about human–robot affective interaction, with in-depth understanding of edge-cloud collaboration mechanisms.
- Strong logical thinking and capable of designing rigorous and reasonable experimental schemes.