TikTok
Machine Learning Engineer Intern (E-Commerce Supply Chain & Logistics-LLM/Agent) - 2027 Start (PhD)
Seattle, Washington · Internship · Posted Oct 28, 2026
- Education
- Currently pursuing a PhD degree in artificial intelligence, computer science, operations research, automation, statistics, mathematics, or a related discipline.
About the job
Develop LLMs and agent systems for global e-commerce supply chain and logistics, including domain modeling, multimodal understanding, workflow automation, and evaluation.
What you'll do
- Build and improve logistics LLMs and intelligent systems, including post-training, reinforcement learning, and inference optimization.
- Develop multimodal and structured understanding for logistics scenarios and integrate capabilities with forecasting and decision systems.
- Build and evaluate agents and workflow tools, including retrieval, orchestration, execution, monitoring, and continuous improvement.
What they ask for
- Currently pursuing a PhD in artificial intelligence, computer science, operations research, automation, statistics, mathematics, or a related discipline.
- Experience with LLMs, agents, RAG, tool use, post-training, evaluation, or applied NLP systems.
- Strong Python programming and familiarity with at least one of Java, C++, Go, or TypeScript.
- Familiarity with machine learning and deep learning frameworks or ecosystems such as PyTorch, TensorFlow, JAX, vLLM, Hugging Face, LangChain, or LlamaIndex.
Nice to have
- Experience with agent systems, LLM post-training, benchmarks, or evaluation datasets.
- Experience in e-commerce, logistics, operations research, data platforms, knowledge graphs, or enterprise knowledge management; relevant publications, open-source work, or projects.
Tools
Python, Java, C++, Go, TypeScript, PyTorch, TensorFlow, JAX
AI focus
Develop and evaluate LLMs, agents, retrieval systems, reinforcement-learning and post-training methods, multimodal logistics understanding, and workflow automation.
Model-written summary · the original posting is authoritative