Machine Learning/ Search Engineer - Services Special Projects

Our team is building a massive, real-time search experience from the ground up - one that will reach users at Apple scale. Its search at the intersection of Generative AI and Information Retrieval, and its a rare opportunity to shape a product that millions will rely on.

We are seeking a highly experienced and innovative Search Systems Engineer to help design, develop, and optimize large-scale search systems.

This role is ideal for a technically deep individual who has a strong product sense and enjoys solving real-world problems using modern AI models and scalable systems. We are a passionate team of hardworking engineers and scientists, and we are looking for a strong Search engineer to join us. You will work closely with AI/ML Scientists and engineers at the intersection of Generative AI and Information Retrieval, crafting intelligent systems that personalize user experiences. Design, build, and maintain large-scale, low-latency, high-performance search systems that can scale. Develop and optimize ranking, relevance, and retrieval through ML/AI models and merging traditional keyword search with vector-based semantic search using embedding models and vector databases. Develop sophisticated NLP pipelines for intent classification, entity extraction, semantic parsing, and query expansion. Merge traditional keyword search (BM25) with vector-based semantic search using embedding models and vector databases. Design and Implement machine learning models (e.g. Learning to Rank, Cross Encoder based models) and multi-stage reranking algorithms to optimize search precision and recall. Build offline and online evaluation metrics, A/B testing frameworks, and continuous improvement strategies for search quality Partner with Research Scientists, Product, Data Engineering, MLOps, Search Infrastructure teams, and UX to align search features with business and user goals. Stay current with the latest research and innovations in search and information retrieval technologies, translating them into scalable production systems.Bachelors or Masters degree in Computer Science, Machine Learning, Statistics, or a related field 10+ years of experience in Machine Learning, Data Science, or Software Engineering roles with a significant focus on search infrastructure and information retrieval. Hands on experience building and deploying large-scale search systems in production. Deep understanding of information retrieval, query understanding, query augmentation and multi-stage ranking algorithms Strong foundation in deep learning architectures for search and retrieval (e.g., transformers, cross encoder models, graph neural networks, learned sparse representations). Experience with to multi-objective optimization in search systems (e.g., relevance, diversity, freshness, fairness). Experience with real-time systems, user feedback loops, and model retraining pipelines. Strong proficiency in Go, Java, C++ and Python Proven experience with ML frameworks including PyTorch, XGBoost. Familiarity with cloud environments (including AWS) and containerization (Docker, Kubernetes) Extensive experience working with data processing pipelines including Spark, Flink Hands-on experience with vector search including FAISS Familiarity with streaming platforms including Apache Kafka Experience with search infrastructure including OpenSearch, and/or Elasticsearch Hands-on experience deploying, serving, and optimizing LLMs, Embeddings and ML models directly in the production query/request path Past successful deployments with tuning of models (including quantization) for performance and quality optimization Excellent communication skills and a collaborative mindsetMasters Degree; PhD Preferred Published work or patents in the domain of search systems, information retrieval, or related ML fields. Experience with graph databases such as TigerGraph Experience with data and model versioning tools and practices (e.g., DVC, MLflow, Weights & Biases) Deep Experience with KV Stores including SSTables and Cassandra Experience with tuning KV-cache and batching for low-latency, high-throughput real-time inference. Deep production level experience with inference runtimes/compilers (ONNX Runtime, TensorRT/TensorRT-LLM), and serving frameworks (vLLM, SGLang or Triton, TorchServe ) .