About the Role:
As a cornerstone of our AI laboratory, you will architect cutting-edge reasoning models and reinforcement learning frameworks to transform complex global data into autonomous, high-precision advertising decisions.
Job Responsibilities
Core Model Development:
Research and develop data-driven AI models, including Deep Learning, Reinforcement Learning (RL), and Large Decision Models. Drive end-to-end excellence in data synthesis, Instruction Fine-Tuning (SFT), and Preference Alignment to enhance model adaptability and quality.
Reasoning & LLM Innovation:
Architect and optimize advanced Reasoning Models, covering the full stack from SFT and RLHF to RAG (Retrieval-Augmented Generation) and intelligent Agentic Workflows.
Business-Impact Integration:
Collaborate closely with Product and Engineering teams to translate complex business requirements into sophisticated AI solutions, significantly improving the efficacy of global ad bidding and decision-making.
Requirements
Academic Background:
Bachelor’s degree or higher in Computer Science or related fields. Solid foundation in Machine Learning and Mathematics with a proven track record in NLP, Computer Vision, or Recommendation/Ad systems.
Technical Expertise:
Specialized in Reinforcement Learning and Large Language Models (LLMs). Publication in top-tier conferences (NeurIPS, ICML, ICLR, etc.) or top placements in prestigious AI competitions is highly preferred.
Agentic Architecture:
Deep understanding of LLM Agent architectures with a demonstrated ability to deploy autonomous Agent applications in real-world scenarios.
Engineering Proficiency:
Mastery of deep learning frameworks such as PyTorch or TensorFlow, with the ability to implement complex algorithms independently from scratch.
Soft Skills:
Highly self-driven, proactive, and optimistic. Excellent communication skills with a strong "Customer-Centric" mindset to solve real-world partner challenges.
Preferred Qualifications (Bonus)
Ad-Tech Domain Knowledge:
Experience in internet advertising, including search/recommendation algorithms, creative asset generation, or digital marketing logic.
LLM Paradigm Mastery:
In-depth knowledge of LLM scaling laws, Fine-tuning, Prompt Engineering, Vector Databases, and the LangChain/LlamaIndex ecosystem.
Proven Impact:
A portfolio of high-impact LLM projects or influential research papers.
Large-Scale Training:
Hands-on experience with pre-training, SFT, and RLHF for 10B/100B+ parameter models from the ground up.
Elite Coding Skills:
Exceptional programming ability; winners of ACM-ICPC, NOI/IOI, or Topcoder are highly regarded.
