Recommendation Algorithm Engineer

Job Overview

Join a stealth Web3 startup in Singapore as a Recommendation Algorithm Engineer responsible for crafting a comprehensive personalized recommendation system tailored for a prediction market platform. This role involves enhancing personalized home feeds, trending events, sports/event markets, and other recommendation contexts, targeting key business metrics such as click-through rate, user engagement duration, participation rates, retention, and conversion in trading.

Responsibilities

• Construct and refine the entire end-to-end personalized recommendation system for the platform's various recommendation scenarios.
• Develop multidimensional user profiles derived from diverse interactions like browsing history, clicks, follows, trading activities, dwell time, and search inputs to capture user interests, risk appetite, thematic tendencies, and engagement levels.
• Design and fine-tune the full recommendation pipeline encompassing cold start handling, candidate retrieval methods, ranking models predicting CTR and CVR, diversity maintenance, multi-objective optimization, real-time relevance adjustments, and trending content promotion.
• Create a scalable framework for feature engineering by analyzing user behavior sequences and extracting high-quality recommendation features.
• Continuously enhance recommendation models and strategies using data-driven iterations, balancing factors such as personalization accuracy, content variety, popularity concentration, and overall user experience. Set up robust evaluation, real-time monitoring, and optimization protocols for recommendation effectiveness.

Qualifications

• Possess at least a Bachelor's degree in Computer Science, AI, Data Science, or related disciplines along with more than 5 years of direct experience building production-grade search or recommendation systems.
• Demonstrate a deep understanding of recommendation system components including candidate retrieval (vector-based, rule-based, popularity-based), ranking, re-ranking, multi-objective optimization, and cold start strategies with proficiency in common CTR/CVR prediction models.
• Be skilled in deep learning platforms such as PyTorch or TensorFlow, capable of independently performing feature engineering, model training, offline evaluation, online tuning, and iterative model development.
• Have solid experience in advanced feature engineering, modeling user behavior sequences, and managing sparse datasets effectively.
• Show strong data analytical aptitude and business insight to diagnose recommendation problems like traffic imbalance, poor long-tail exposure, ineffective cold start, and content homogenization, deploying optimized solutions.
• Be familiar with A/B testing frameworks and experimental design approaches to quantitatively assess recommendation strategies and drive improvements.

Preferred Qualifications

• Prior experience with recommendation systems related to financial markets, trading platforms, market data, trending events, news feeds, or sports content is highly advantageous.
• Expertise in vector-based retrieval approaches (such as Approximate Nearest Neighbor and embedding search), personalized ranking methodologies, real-time trending ranking, and multi-objective recommendation modeling.
• Experience in designing user profiling and tagging systems that effectively integrate both short-term and long-term user interest modeling for recommendation tactics.