LLM & Agent Algorithm Expert - TikTok Search
Core
Lead R&D of search-domain LLM foundations and Agent execution frameworks to build a proactive, task-completing search engine for TikTok.
Role type
Senior IC LLM & Agent Algorithm Expert
Builds
Search LLM foundation, Agent execution framework (Harness), multi-agent collaboration systems, and self-improvement loops.
Domain
Search, Large Language Models, Multi-Agent Systems, Multimodal AI
Deliverable
production ML models
Required skills
LLM pre-training/post-training, Agent orchestration, multi-agent scheduling, inference optimization, distributed training, data synthesis, causal reasoning, automated evaluation
Preferred skills
Top-tier conference publications, open-source leadership, large-scale business implementation
Technologies
TensorRT, MapReduce, Spark, Harness, RLHF, SFT, quantization, pruning, distillation
Responsibilities
Own pre-training/post-training pipelines for search LLMs; design multi-agent cluster scheduling and collaboration algorithms; build long-term memory and self-improvement mechanisms; develop data synthesis and automated evaluation systems; implement long-horizon task Agents and multimodal AIGC features.
Seniority
Senior, hands-on IC
