多媒体AI算法工程师-TikTok
Core
Develop multimodal (image, video, audio) invisible watermark and digital fingerprint extraction technologies, optimizing model robustness and reliability.
Role type
Senior IC multimodal AI algorithm engineer
Builds
Invisible watermarking and digital fingerprinting services and SDKs
Domain
Multimedia security and digital forensics
Deliverable
production ML models
Required skills
Deep learning model training and optimization, model quantization, CPU-GPU co-optimization, Python, C++, Go
Preferred skills
Audio/video processing, digital signal processing (DSP), generative models (Stable Diffusion, DiT, LLMs), security research, top-tier conference publications (CVPR, ICCV, NeurIPS)
Responsibilities
Train and optimize deep learning models for watermark extraction, develop services and SDKs, establish evaluation metrics, track and reproduce cutting-edge academic research
Seniority
Senior, hands-on IC
