ML Architect

Job Title

Responsibilities:

Design hybrid retrieval systems combining keyword search, vector similarity, and cross-encoder reranking at scale. Build intelligent query routing with cascading classification strategies. Architect multi-model inference pipelines optimized for latency-sensitive workloads. Define relevance metrics, run A/B experiments, and drive measurable business outcomes. Support the driving MLOps standards for model deployment, monitoring, and continuous improvement. Partner with Product, Merchandising, and Engineering to translate business requirements into ML solutions. Mentor engineers and define search and ML architectural standards.

Requirements:

7+ years in software, data, or ML engineering with 3+ years building production search systems. Experience with e-commerce search patterns: faceting, merchandising rules, query understanding. Strong knowledge of embedding models, approximate nearest neighbor search, and reranking architectures. Hands-on experience with vector databases and similarity search at scale (Pinecone, Milvus, Weaviate, FAISS or similar). MLOps expertise: model deployment pipelines, monitoring, versioning, and retraining workflows. Production experience with transformer-based models for classification and ranking. Track record balancing latency, cost, and relevance tradeoffs in real-time systems.