LLM & Agent AI Research Intern

TikTok; Singapore; On-site; SGD 17,000 - 28,000; Full time

Job summary

TikTok is offering a PhD internship in the TT-Search Algorithm & Applied AI team in Singapore. You will contribute to building the search-domain LLM foundation, agent orchestration, memory/self-improvement loops, and evaluation frameworks as part of a dynamic internship program.

Ideal candidates have a PhD candidacy, strong ML/DL grounding, and experience with LLM/Agent projects, PyTorch/TensorFlow, and big data tools. Applications reviewed on rolling basis with early availability preferred.

Qualifications

• PhD candidate in Computer Science or related field.
• Strong ML/DL foundation and familiarity with LLM/Agent technologies.
• Experience with LLM pre-training/fine-tuning/alignment and agent development is preferred.
• Familiarity with PyTorch/TensorFlow for model training and deployment; knowledge of distributed training and inference acceleration.

Responsibilities

• Search LLM Foundation R&D: build and optimize the search-domain LLM foundation and post-training pipelines.
• Agent Engineering & Multi-Agent Orchestration: develop the Harness framework and multi-agent coordination algorithms.
• Long-term memory, self-improvement and RSI: explore memory mechanisms and automated optimization pipelines.
• Evaluation & Reward/Verifier Systems: contribute to online evaluation and annotation-free evaluation systems.
• Business Implementation: work on multimodal AIGC, visual search, on-device intelligence and related tasks.

Skills

ML/DL foundation

LLM/Agent experience

PyTorch/TensorFlow

Big data (MapReduce/Spark)

Education

PhD in Computer Science or related

Tools

PyTorch

TensorFlow

MapReduce/Spark

Job description

TikTok is offering a PhD internship in the TT-Search Algorithm & Applied AI team in Singapore. You will contribute to building the search-domain LLM foundation, agent orchestration, memory/self-improvement loops, and evaluation frameworks as part of a dynamic internship program.

Ideal candidates have a PhD candidacy, strong ML/DL grounding, and experience with LLM/Agent projects, PyTorch/TensorFlow, and big data tools. Applications reviewed on rolling basis with early availability preferred.