Applied Scientist, Core Search

The Amazon Search team"s vision is to deliver high quality search results regardless of how customers phrase their search queries. Keyword-based search breaks down when confronted with natural language expressions. Queries like "I have ants in my house," "headphones comparable to Bose," "breakfast foods for someone avoiding sugar," and "scratch resistant flooring for dogs that looks like real wood" require world knowledge, common-sense reasoning, and sophisticated language understanding that customers increasingly expect.

Core Search team is reimagining search architecture using Large Language Models (LLMs): a new LLM stack that already powers Amazon Search, Alexa+, Alexa for Shopping, Help Me Decide, Interests AI, confidential initiatives, and a growing portfolio of Amazon experiences across Stores and Devices. We build this stack as a primitive to supercharge a new generation of natural-language experiences across Amazon.

We are hiring an Applied Scientist to push the science behind this stack: the reasoning LLMs, embedding models, cross-encoder rankers, and multi-objective optimization systems that turn billions of products into the right answer for hundreds of millions of customers. The role spans the full model lifecycle, from mid-training reasoning models on shopping data to aligning the models with customers on the dimensions that matter for shopping: helpfulness, trust, and faithfulness.

You will build with us a natural language AI interface to billions of products, for all Amazon customers.

Key job responsibilities

As an Applied Scientist on the team, you will lead science innovation across multiple problems and surfaces. You will:

• Develop personalized multi-modal thinking-LLM techniques that reason about customers, queries, and products.
• Mid-train and post-train large language models on shopping data: domain-adaptive continued pre-training, ireinforcement learning shopping reasoning traces, and instruction tuning for natural-language shopping queries.
• Align models with customer interests on the dimensions such as helpfulness, harmlessness, and faithfulness. Apply Reinforcement Learning (RLVR, RLHF), Direct Preference Optimization (DPO), and customer-behavior-derived reward models.
• Create semantic representations of products, customers, and context (bi-encoder embeddings, contrastive learning, hard-negative mining, cross-lingual training).
• Develop cross-attentive LLM rankers that score candidate products against rich query intent and complex constraints.
• Train multi-objective ranking and optimization systems that balance relevance, purchasability, and personalization.
• Drive improvements on offline benchmarks as well as online experiments.

About the team

Core Search builds the next-generation LLM-powered retrieval and ranking stack for Amazon. We own the stack end-to-end including LLM models, personalization, multi-turn natural-language refinements, routing, the experimentation service, and the partner-facing primitive that other Amazon teams build on top of. The team is highly motivated, collaborative, technically deep, and runs with strong executive sponsorship and strategic visibility. In this role, you will define program strategy, prioritize investments, and shape how AI-driven natural-language search experiences ship across all devices, globally.