Machine Learning Engineer Graduate (Conversational AI) - 2027 Start
Core
Building the next-generation unified Agent runtime and self-evolving loop for TikTok's global e-Commerce customer service across 30+ languages.
Role type
Senior IC machine-learning engineer (agent runtime & evaluation systems)
Builds
Self-improving agent runtime, context/memory systems, post-training pipelines, and evaluation infrastructure for live commerce.
Domain
E-Commerce, Conversational AI, Large Language Models (LLMs)
Deliverable
production ML models
Required skills
Python, C++/Go/Rust/Java, Machine Learning, Deep Learning, NLP, Statistical Literacy, LLMs, Agents
Preferred skills
Agent runtime/harness development, Post-training (SFT/DPO/RLHF), Evaluation (LLM-as-judge), Inference optimization (vLLM/TensorRT-LLM), Multilingual NLP
Technologies
vLLM, TensorRT-LLM, MoE, MCP, SFT, DPO, RLHF, RLAIF, RLVR
Responsibilities
Orchestrate agent skills, tools, and context with safety guardrails; manage agentic memory and context compaction; implement post-training and data flywheel loops; build tool libraries and skill governance; design rigorous evaluation systems; construct self-evolving case mining and candidate generation loops; run online experimentation and causal readouts.
Seniority
Graduate (entry-level with strong fundamentals) (via careerplan.io/jobs/7670685495206857013-machine-learning-engineer-graduate-conversational-ai-2027-start-at-tiktok)
