Conversational AI Researcher for Large Language Models in Global E-Commerce
Singapore · Full Time
Be the first to apply
- Experience
- Any
- Salary
- —
- Openings
- 1
- Posted
- 2 weeks ago
- Work mode
- In office
- Education
- PhD
- Resume
- Required to apply
Where you'll work
Sign in to tell us what does and doesn't work for you here — it sharpens every match we show you.
Job description
Overview
This role focuses on advancing a cutting-edge large language model (LLM) dialogue system designed to serve hundreds of millions of daily users within the Douyin e-commerce ecosystem. The system supports a broad spectrum of customer service functionalities, including platform, merchant, influencer services, and innovative intelligent shopping guidance, aiming to significantly enhance operational efficiency and user experience through continuous technological refinement.
Research Objectives
Develop an LLM-powered chatbot for TikTok and Douyin e-commerce that manages complete user inquiries: clarifying requests, negotiating solutions, and executing responses to elevate intelligent customer service interactions.
Research Approach
Engineer a multi-agent LLM architecture comprising three specialized agents:
- Reply-agent: Ensures compliance with platform policies, prevents hallucinations, and negotiates effectively with users.
- Planning-agent: Identifies user needs and context, sourcing applicable guidelines and recognizing risk scenarios.
- Tool-agent: Validates tools' legal use, interprets tool outputs accurately, and manages dependencies among actions.
Research Challenges
- Ensuring chatbot responses strictly adhere to service policies (e.g., refund timelines, user coupon limits).
- Enabling dynamic adaptation through real-time interaction, personalized responses, and progressive service to match human-level flexibility.
- Incorporating self-reflection for the LLM to analyze and refine its behavior, improving decision-making on complex queries.
- Processing complex images such as shipping photos, bank transactions, damaged goods pictures, and seller certifications to enhance problem resolution.
Qualifications
- PhD degree in computer science, software engineering, AI, mathematics, or related disciplines preferred.
- Strong foundation in machine learning, deep learning, generative and multimodal models, with robust mathematical skills and aptitude for self-learning.
- Proficient in ML frameworks such as PyTorch and TensorFlow and skilled in software engineering and coding.
- Experience with multimodal large models, especially in long-text or film/TV drama domains, is advantageous.
- Preference for candidates with publications in top-tier CS conferences/journals or notable competition experience.
About TikTok
TikTok is a global leader in short-form mobile video, aiming to inspire creativity and bring joy. With headquarters in Los Angeles and Singapore and offices worldwide, TikTok fosters a diverse workforce dedicated to innovation and impact.
Work Culture & Inclusion
TikTok values curiosity, humility, and resilience, embracing challenges with an "Always Day 1" mindset to achieve breakthroughs collectively. The company is passionate about building an inclusive environment that celebrates diverse backgrounds, perspectives, and experiences, reflecting the global community it serves.
Minimum education
Doctorate
Industry
E-Commerce