LLM & Agent Algorithm Project Intern (Search) - 2026 Start (PhD)
Core
Building the search-domain LLM foundation and Agent execution framework (Harness) to shift search from retrieval-and-ranking to proactive task completion.
Role type
PhD-level research intern (LLM & Agent algorithms)
Builds
Search LLMs, multi-agent orchestration systems, self-improvement loops, and evaluation frameworks for TikTok Search
Domain
Search, Large Language Models, Multi-Agent Systems, Multimodal AI
Deliverable
production ML models
Required skills
Machine learning, deep learning, LLM pre-training/fine-tuning/alignment, Agent development (tool calling, planning, memory), RAG, RLHF/RLAIF, data synthesis, automated evaluation, PyTorch, distributed training, model compression/inference acceleration
Preferred skills
Publications at NeurIPS/ICML/ICLR/ACL/EMNLP/CVPR/ICCV/ECCV/AAAI, Kaggle/COCO/ImageNet/ActivityNet awards, CV/NLP domain expertise, graph neural networks, multilingual learning
Technologies
PyTorch, TensorFlow, MapReduce, Spark, TensorRT, TinyLLM
Responsibilities
Develop post-training pipelines for search LLMs (ultra-long-text, multimodal), build multi-agent cluster scheduling and collaboration algorithms, work on long-term memory and self-evolution mechanisms, construct online Reward/Verifier systems and automated evaluation frameworks
Seniority
PhD candidate, research-focused
