Senior Research Scientist/Software Engineer, LLM/Agent Platform (TikTok-Content Ecology AI Innovation & Platform)
Core
Design and implement novel LLM/Agent frameworks with self-improving capabilities and architect robust, high-performance infrastructure for global scale.
Role type
Senior Research Scientist / Architect (LLM/Agent Platform)
Builds
Scalable infrastructure for LLMs and agents, including model registries, feature stores, and high-throughput retrieval/RAG systems.
Domain
Artificial Intelligence, Large Language Models, Agent Systems, Distributed Systems
Deliverable
production ML models
Required skills
Deep learning frameworks (PyTorch, Tensorflow, JAX/FLAX), distributed systems, model lifecycle management, observability frameworks, evaluation frameworks, knowledge discovery workflows
Preferred skills
GPU scheduling, model serving (Triton, TensorRT-LLM), vector databases (FAISS, Milvus), MLOps, distributed training, KV cache–aware serving, GPU/accelerator optimization, high-performance networking (RDMA, NCCL), performance optimization (DeepSpeed/ZeRO, vLLM), agent architectures (planning, tool use, memory)
Technologies
PyTorch, Tensorflow, JAX/FLAX, Triton, TensorRT-LLM, FAISS, Milvus, DeepSpeed, ZeRO, vLLM, LangChain, LlamaIndex, RDMA, NCCL
Responsibilities
Architect standardized pipelines for model/agent lifecycle from data processing to deployment; Partner with algorithm teams to provide infrastructure accelerating R&D; Build observability and evaluation frameworks for reproducibility and cost-efficiency; Design core platform infrastructure including registries and feature stores; Design and optimize Agentic AI systems with planning, tool use, and memory components.
Seniority
Senior, hands-on IC with architectural responsibilities
