Staff Applied AI Researcher (Agentic Search)

Role overview Set applied research direction for an agent-oriented search system that retrieves and evaluates live web information for AI applications. The role spans research and production engineering: you will design retrieval and ranking methods for both quick queries and multi-step searches, then ship improvements under demanding latency, reliability, and scale constraints. In the first year, you are expected to take ownership of a defined part of the search stack, such as ranking or iterative retrieval.

Responsibilities - Guide technical research across search, retrieval, and ranking, and evolve multi-stage systems for understanding, rewriting, retrieving, reranking, and refining queries. - Experiment with embedding approaches, sparse and hybrid retrieval, reranking, and context compression, then bring successful methods into production. - Develop ways to ground language models in current web data, accounting for source trustworthiness, quality, and freshness. - Work with evaluation specialists to define metrics for agentic search, where success cannot be measured by clicks alone. - Analyze relevance, latency, and cost trade-offs and partner with engineers to deploy reliable systems at web scale. - Mentor researchers and engineers and help set a high technical standard for the team.

Requirements - 8+ years of experience in applied AI, machine learning, or software engineering. - A track record of shipping ML or AI systems into production at scale. - Deep experience in search, retrieval, ranking, or retrieval-augmented assistants. - Strong understanding of modern deep learning, especially transformers and embeddings, and experience with language-model-integrated or knowledge-intensive systems. - Experience designing evaluation frameworks and metrics for machine learning systems; strong Python programming skills. - Research publications at leading venues such as ICLR, ICML, NeurIPS, AAAI, EMNLP, or ACL, and ability to work with autonomy in a fast-moving product environment.

Nice to have Experience with agentic AI, tool use, or autonomous workflows; reinforcement learning post-training for language or vision-language models.