Search &RecommendationAlgorithmEngineer

Responsibilities

1. End-to-End Search Algorithm R&D and Optimization

• Design and iterate the algorithm system for overseas large-screen content search, covering user intent understanding, multilingual query parsing, multi-path recall, fine ranking, and re-ranking.
• Continuously improve core metrics, including search CTR, play conversion rate, and search satisfaction.

2. End-to-End Personalized Recommendation Algorithm Development

• Build a dual-layer user profile system that distinguishes the household group from individual members.
• Own end-to-end algorithm design and iteration across the recommendation pipeline: recall, coarse ranking, fine ranking, and re-ranking.
• Develop personalized recommendation models for home viewing and leisure scenarios. These models should improve content matching accuracy, recommendation CTR, session duration, play depth, and user engagement.
• Mine users’ long-term and short-term interests, and solve recommendation challenges such as conflicting interests among family members and cold start.

3. Unified Search & Recommendation Architecture

• Lead the design and implementation of an integrated architecture for search and real-time recommendation.
• Unify content representation, user modeling, and ranking systems so that search intent and personalized recommendation work together closely.
• Enable two-way linkage: search behavior feeds into recommendation, and search intent can be triggered from within recommendation scenarios.

4. Global Multilingual Search & Recommendation Capabilities

• Optimize cross-lingual semantic retrieval, multilingual intent understanding, and unified content vector representation for global markets.
• Ensure accurate, consistent, and intelligent results for both voice and text search.
• Support personalized recommendation in multilingual scenarios.

5. Data Closed Loop & Experimentation

• Work with product, data, and engineering teams to build a full data closed loop for both search and recommendation. The loop should cover exposure, clicks, conversions, and negative feedback.
• Continuously validate strategies through A/B testing to iterate algorithm models and business strategies.

Basic Qualifications

• Bachelor’s degree or above in Computer Science, Mathematics, Statistics, Software Engineering, or a related field. A Master’s degree is preferred.
• 5+ years of industry R&D experience in search, recommendation, or NLP algorithms.
• Proven experience deploying and optimizing large-scale online search or recommendation systems end to end. Experience in joint search-recommendation modeling is a plus.

Algorithm & Engineering Skills

• Solid foundation in machine learning and deep learning. You should be familiar with mainstream search and recommendation models, such as two-tower recall, DeepFM, Wide & Deep, DIN, and multi-objective ranking.
• Proficiency in Python, with hands-on experience in at least one deep learning framework (PyTorch or TensorFlow).
• Familiarity with search engines and vector databases such as Elasticsearch, Faiss, and Milvus, including online tuning and deployment.
• Familiarity with feature engineering, user and content profile systems, and real-time and offline computing frameworks (Flink, Spark). Experience in recommendation features and user behavior sequence modeling is expected.

Business & Experimentation Skills

• Familiarity with core search and recommendation metrics (CTR, CVR, Recall, NDCG, etc.) and A/B testing methodology.
• Strong data analysis, problem diagnosis, and root-cause analysis skills.
• Hands-on experience with large-scale online search and recommendation systems, including independently leading module iteration and performance optimization.

Preferred Qualifications

• Algorithm experience in search and recommendation for overseas large-screen devices, smart home terminals, or streaming content is a strong plus.
• Experience with multilingual NLP, cross-lingual retrieval, or global search and recommendation systems.
• Experience in family multi-user modeling, scenario-aware personalized recommendation, LLM retrieval augmentation (RAG), or unified search-recommendation modeling.