Algorithm Engineer-TikTok E-Commerce (Conversational AI)

proving, providing more accurate, context-aware, and human-like interactions.
By joining us, you will be at the forefront of transforming customer service in e-commerce, helping build an AI system that understands, adapts, and provides intelligent solutions-all while reducing costs and improving efficiency for merchants and the platform.
What You Will Do

• Develop AI-Powered Customer Service Systems: Design and implement an AI-driven conversational customer service agent that can handle e-commerce inquiries, complaints, refunds, dispute resolutions, and logistical issues, replacing traditional human customer service agents.
• LLM Post-Training & Data-Efficient Learning: Apply state-of-the-art LLM post-training techniques, such as instruction tuning, reinforcement learning from human feedback (RLHF), and continual learning, to optimize AI customer service responses with minimal labeled data.
• Benchmark and training data construction: Identify challenging customer service interactions, such as policy clarifications, dispute handling, and multi-turn complaint resolution, and construct specialized datasets to enhance AI training.
• Develop Multilingual Customer Support: Build AI models capable of handling customer service interactions across multiple languages and cultural contexts, ensuring accurate translation and appropriate responses for a diverse global audience.
• Optimize Model Efficiency & Deployment: Work on model compression, quantization, and efficient inference techniques to ensure the AI customer service assistant can run at scale with low latency and high reliability.
  Responsibilities

1. Develop AI Customer Support Systems: Build and optimize AI-driven customer service models capable of handling high-volume, complex user inquiries while ensuring high response accuracy and reliability.
2. Enhance LLM-Based Customer Interaction Models: Implement LLM post-training strategies to improve customer support interactions, reducing errors, hallucinations, and irrelevant responses.
3. Create Automated Dispute Resolution & Policy-Adaptive AI: Develop intelligent models capable of handling disputes, verifying transaction details, and ensuring platform compliance in automated responses.
4. Develop Multilingual Support & Translation Models: Enhance the platform's AI translation capabilities for real-time multilingual customer service interactions, ensuring smooth cross-language communication.
5. Refine Response Evaluation Metrics: Define and implement quality evaluation metrics for AI-generated responses to track customer satisfaction and improve conversational AI quality through A/B testing and iterative optimization.
6. Enable AI-Seller Collaboration: Build AI-powered seller assistance tools to help merchants quickly respond to customer inquiries, manage store operations, and resolve disputes efficiently.
7. Optimize Large-Scale Model Deployment: Work on model compression, inference optimization, and edge AI deployment to ensure real-time, high-quality customer service experiences at scale.

Qualifications

Minimum Qualifications
-Bachelor and above with majors in computer science, computer engineering, statistics, applied mathematics, data science or other related disciplines.

• At least 3 years of work experience in the related field
• LLM Development & Post-Training Expertise: Experience in fine-tuning, distillation, or reinforcement learning of large language models for conversational AI applications.
• Multilingual AI Development: Proficiency in multilingual NLP, machine translation, and cross-lingual dialogue modeling.
  Preferred Qualifications
• E-commerce Business Acumen: Understanding of e-commerce policies, dispute resolution workflows, and merchant-buyer interactions to enhance AI service design.
• Advanced NLP & Deep Learning: Strong grasp of AI agents, retrieval augmented generation, mixture of experts, sparse attention, reinforcement learning, inference time scaling etc. for improving AI dialogue quality.
• Scalability & Efficiency: Experience in distributed model training, low-latency inference, and edge AI for large-scale customer service applications.