Senior Software Engineer, Computer Vision (Knowledge Distillation)

The Vision Engineering Team at TikTok is at the forefront of delivering GenAI technologies directly into the TikTok products worldwide. Leveraging our proprietary AI infrastructures, we streamline the creation, integration, testing, and deployment of the GenAI features. This also encompasses large-scale training stability and optimization for acceleration, as well as large model inference and multi-machine multi-card deployment. Our work enhances user experience by powering diverse functionalities, including visual enhancements, video editing tools, and creative camera filters, both within TikTok and other applications.

The team is looking for strong software engineers who can take initiatives to design and implement highly efficient engineering systems for generative AI tasks, including but not limited to model training and optimization, model deployment with efficient hardware consumption, and applications such as image/video processing and interactive editing.

Responsibilities
- Develop algorithm acceleration technologies for text-to-image/text-to-video models through knowledge distillation, model architecture redesign (dynamic MoE routing/sparse attention), and parameter-efficient design (low-bit quantization) to achieve order-of-magnitude efficiency gains.
- Lead generative model innovation with focus on diffusion acceleration (sampling step reduction/latent optimization), autoregression model efficiency.

Minimum Qualifications:
- B.S./M.S./Ph.D. in Computer Science or related fields with 5+ years of software development experience in C++/Python.
- Expertise in diffusion models (Stable Diffusion/DiT) with deep understanding of computational bottlenecks and optimization methodologies.
- Proven experience in ≥1 domain: Image/Video Generation, Image/Video Editing/Understanding, Model Compression (Quantization/Knowledge Distillation), Efficient Architectures (MoE/Sparse attention), Generative Alignment (RLHF/DPO).
- Excellent communication and teamwork skills, capable of thriving in a fast-paced work environment.


Preferred Qualifications
- Kaggle competition achievements, publications at ICML/NeurIPS/CVPR, or open-source contributions (e.g., HuggingFace Diffusers optimization).
- Research experience in GenAI /MLsys areas.
- Familiarity with open source deep learning frameworks such as Pytorch/DeepSpeed/Jax etc.