返回查询:Machine Learning / 广州市

We are looking for individuals who are passionate about
pushing the boundaries of computer vision
while also delivering efficient, production-ready models. Candidates with a strong record of open-source contributions or a demonstrable portfolio of impactful ML projects will stand out.

Key Responsibilities

  • Design, implement, and optimize image/video ML models for large-scale generative AI applications
  • Apply advanced techniques such as quantization, pruning, fine-tuning, and model distillation to improve efficiency and performance
  • Work with deployment frameworks (ONNX, TensorRT, CUDA) to enable scalable inference
  • Collaborate with researchers to bridge state-of-the-art techniques with production-grade engineering
  • Conduct experiments, benchmark models, and document findings
  • Share insights and collaborate with the broader ML research/engineering team

Required Qualifications

  • Bachelor's, Master's, or PhD in Computer Science, AI, Machine Learning, or related fields
  • Hands-on experience with deep learning frameworks: PyTorch (preferred), TensorFlow, or JAX
  • Strong background in optimization techniques (quantization, pruning, distillation, mixed precision)
  • Experience with deployment toolkits: ONNX, TensorRT, CUDA, or similar
  • Demonstrated work in image/video model training, fine-tuning, or adaptation
  • Proficiency in Python and familiarity with large-scale GPU-based training

Preferred Qualifications

  • Experience with GGUF, ComfyUI, or other model deployment ecosystems
  • Contributions to open-source computer vision or ML frameworks
  • Research or industry experience with foundational/generative models (diffusion, transformers, multimodal)
  • Familiarity with scaling ML models in distributed systems

Portfolio / Contribution Requirement

  • Please share your GitHub profile showcasing open-source work in ML/Computer Vision
  • If unavailable, provide a portfolio of projects, research papers, or documented achievements in model development, training, or optimization

Powered by JazzHR

T1TzpQ8rsQ