Senior Staff Algorithm Engineer Recommendation
OKX is a leading global cryptocurrency exchange offering a wide range of trading services, including spot and derivatives trading, as well as Web3 solutions.
Investors
Projects
About OKX
OKX is a global cryptocurrency exchange known for its trading services and financial products. The platform offers a wide range of features, including spot and derivatives trading, staking services, and a user-friendly interface suitable for both beginners and experienced traders.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will own the technical direction of a next-generation social feed recommendation system. You will improve ranking models, build unified user-intent representations, define the technical roadmap toward generative recommendation, and integrate recommendation and search capabilities into an LLM Agent framework.
Requirements
- Master's degree or above in Computer Science or Mathematics
- 8+ years of experience, including 5+ years in recommendation or search roles
- Experience owning end-to-end recommendation pipelines at 10M+ DAU scale
- Experience designing unified intent representations across content, feature, and search domains
- Experience with real-time behavioral signals and long-term user preferences
- Deep understanding of attention mechanisms and sequential behavior modeling
- Proficiency in ListMLE, Softmax Loss, and joint multi-candidate ranking
- Expert knowledge of MMoE, PLE, ESMM, gradient conflict mitigation, and composite loss functions
- Experience connecting offline metrics such as AUC and NDCG with online business KPIs
- Experience with Uplift Modeling, bias correction, and probability calibration
- Knowledge of semantic tokenization, conditional sequence generation, RLHF, or DPO
- Experience with LLM Agent frameworks, tool use, ReAct, and natural-language intent translation
- Experience with large-scale distributed training, Flink, Kafka, and inference optimization
Responsibilities
- Elevate the ranking system through continuous model iteration
- Build a cross-domain user-intent framework
- Define and execute a 12–24 month technical roadmap toward generative recommendation
- Integrate recommendation and search capabilities into an LLM Agent framework
Benefits
- L&D programs
- Education subsidy
- Team-building programs
- Company events
- Wellness allowance
- Meal allowance
- Comprehensive healthcare schemes for employees and dependants