AI Infrastructure Engineer - Recommendation & LLM

About the Team We are looking for experienced Software Engineers / ML Systems Engineers to join our Model Infrastructure team and build the next generation of AI infrastructure powering TikTok's For You recommendation system and Large Language Models (LLMs). Our team develops the core training and serving infrastructure behind one of the world's largest recommendation systems, enabling billions of personalized recommendations every day. We are also building next-generation infrastructure for foundation models and LLMs, covering large-scale model training, online inference, GPU optimization, distributed systems, and AI serving. As a member of the team, you will take ownership of challenging infrastructure problems at massive scale, working across model, framework, runtime, GPU, distributed systems, and production serving. You will collaborate closely with researchers, algorithm engineers, and infrastructure teams to turn state-of-the-art AI technologies into highly scalable, reliable, and efficient production systems. This role is ideal for engineers who are passionate about AI systems, distributed computing, GPU optimization, Recommendation, LLMs, and building high-performance infrastructure at massive scale.

Responsibilities - Design, develop, and optimize large-scale AI training and online inference infrastructure for recommendation models and LLMs. - Drive the architecture and implementation of distributed training and serving systems with high scalability, reliability, and efficiency. - Optimize end-to-end model performance across GPU computation, communication, memory, networking, and runtime systems. - Develop and optimize LLM training and serving infrastructure, including model parallelism, KV Cache, Continuous Batching, and efficient inference. - Work closely with researchers and algorithm engineers to productionize new model architectures and algorithms. - Identify and resolve performance bottlenecks across the full AI stack, from model and framework to GPU kernels and distributed runtime. - Improve system latency, throughput, GPU utilization, scalability, and infrastructure cost efficiency. - Drive technical design, implementation, performance benchmarking, and production rollout of critical infrastructure components. - Mentor junior engineers and contribute to the team's technical direction and engineering standards.