Staff Applied AI/ML Scientist, Faire

• 7+ years of experience building large-scale AI/ML systems, with 3+ years in recommender systems, search, or ads ranking
• Hands-on expertise with deep learning libraries (e.g. PyTorch), retrieval infrastructures (e.g. ElasticSearch, Faiss, ScaNN, Pinecone), and real-time data processing feature infrastructures (e.g. Kafka, Flink)
• A proven track record of shipping STOA ML models that blend large language models (e.g., BERT, GPT-class) to power personalization
• A product mindset with a bias for action—you’re comfortable going from paper to prototype to production in days with a deep understanding of user problems
• Strong coding skills in Python, good engineering sense to code in production environments, a deep respect for reliability and ownership, and experience working in high-stakes production environments
• Clear communication and a history of cross-functional impact—you influence beyond your team and help set the technical bar

Desirable

• Own the next-gen recommendation engine, combining multi-stage ranking, LLMs, deep personalization embeddings, reinforcement learning, and generative recommendation to serve personalized product recommendations (e.g. feed) in <100ms latency
• Design and productionize agentic systems for discovery, enabling AI agents to generate personalized product explorations, provide explainable recommendations, and assist retailers in browsing, filtering, and evaluating products and brands
• Drive technical strategies to advance various key business priorities, such as cold-start recommendations and long-term value (LTV) optimization, working across Discovery to help retailers find the right brands and cultivate lasting relationships on Faire
• Lead end-to-end model development and deployment efforts and experimentation, working with best-in-class frameworks to scale inference reliably
• Mentor and grow other senior Applied Scientists and MLEs, and establish best practices around deep learning model development, LLM / agent workflow evaluation, and MLOps