AI Agent Architect: Multi-Turn LLM Systems

TIKTOK PTE. LTD.; Singapore; On-site; SGD 180,000 - 280,000; Full time

Job summary

TikTok PTE. LTD. is seeking talented researchers to join the Global E-Commerce Affiliate Algorithm team.

You will contribute to building AI-native workflows, intelligent search and recommendation, and creator/merchant tooling across multiple languages. The role emphasizes hands-on work with LLMs, agent systems, and scalable infrastructure. Successful candidates will tackle bold ideas, engage in end-to-end system development, and participate in a fast-growing tech environment.

Qualifications

• PhD in CS/AI/Math or a quantitative field.
• Strong Python plus one of C++ / Go / Rust / Java.
• Solid ML / DL / NLP fundamentals with hands-on experience with LLMs or agents.
• Able to read a paper or an engineering blog and turn it into working code.

Responsibilities

• Build and evolve the agent runtime powering creator/seller agents — orchestrate skills, tools, and context.
• Engineer context & memory for long multi-turn agents — agentic memory, context curation, and retrieval.
• Post-train and fine-tune LLMs and build data flywheels for training/eval/retrieval signals.
• Design and integrate tools, Skills, and MCP connectors (tools-as-APIs).
• Build evaluation — LLM-as-judge with calibration; manage offline/online gap.

Skills

Python

C++/Go/Rust/Java

ML/NLP fundamentals

Code from papers

Education

PhD in CS/AI/Math or quantitative field

Tools

PyTorch

TensorFlow

Linux tooling

Job description

TikTok PTE. LTD. is seeking talented researchers to join the Global E-Commerce Affiliate Algorithm team.

You will contribute to building AI-native workflows, intelligent search and recommendation, and creator/merchant tooling across multiple languages. The role emphasizes hands-on work with LLMs, agent systems, and scalable infrastructure. Successful candidates will tackle bold ideas, engage in end-to-end system development, and participate in a fast-growing tech environment.