Token Metrics Jobs (NOW HIRING)

Amazon Advertising is one of Amazon's fastest growing and most profitable businesses, responsible for defining and delivering AI-powered solutions that transform how advertisers make strategic decisions. We deliver billions of ad impressions and process massive volumes of advertiser data every single day. You'll work with us to pioneer breakthrough approaches in how AI agents access and reason over real-time advertiser data at scale.
We are using generative AI and agentic systems to help advertising agents provide instant, strategic advice to millions of advertisers

You will need to invent new techniques for agent orchestration, context optimization, and code generation to ensure we're delivering accurate, trustworthy insights with minimal latency and token consumption. You'll create feedback loops to ensure our solutions are constantly evaluating themselves and improving.
The Ads Real-Time Data Service team is seeking an exceptional Applied Scientist to research and develop novel approaches for agent-data interaction. The Ads Real-Time Data Service team is solving one of the most critical challenges in advertising AI: instant access to advertiser context

We're building the infrastructure that provides immediate, pre-computed access to advertiser data via Model Context Protocol (MCP) servers-an emerging standard for AI agent-data interaction. We're building summarized data for context using a mix of state of the art techniques like CodeAct and RAG-based embeddings, achieving a fundamental transformation in how AI agents interact with data.
This role balances applied research (60%) with productionization (40%), giving you the opportunity to both advance the state of the art and see your innovations deployed at Amazon scale.
Key job responsibilities
Agent Orchestration & Optimization Research
- Research and develop novel algorithms for agent-data interaction patterns that minimize latency, token consumption, and error rates
- Investigate multi-agent orchestration strategies for complex advertiser queries requiring data from multiple sources
- Develop techniques for automatic query optimization and caching strategies based on agent behavior patterns
Large Language Model Context & Token Optimization
- Invent new methods for compressing advertiser context representations while preserving semantic meaning and analytical utility
- Research optimal metadata generation techniques that help large language models understand and reason over structured advertiser data
- Design evaluations to measure the impact of different data representations on agent response quality and token efficiency
- Develop adaptive context selection algorithms that dynamically choose relevant data based on query intent
RAG-Based Embeddings & Semantic Search
- Pioneer new RAG-based embedding approaches optimized for real-time advertiser data delivery with sub-second latency
- Research and implement semantic search and retrieval techniques for advertiser datasets using vector embeddings
- Design advertiser context frameworks that enable automatic schema mapping from advertiser concepts to data representations
- Develop evaluation frameworks to measure performance across dimensions of latency, accuracy, and developer experience
Experimentation & Productionization
- Design and execute rigorous experiments comparing traditional API orchestration versus CodeAct patterns and RAG-based approaches across metrics like success rate, latency, token consumption, and response quality
- Analyze large-scale advertiser interaction data to identify patterns, bottlenecks, and optimization opportunities
- Collaborate with engineering teams to productionize research innovations and deploy them to 30+ advertising agents and skills
- Establish evaluation metrics and benchmarks for agent-data interaction performance
Cross-Functional Collaboration & Thought Leadership
- Partner with agent builder teams to understand their data requirements and constraints
- Work with platform engineers to implement and optimize MCP servers, data pipelines, and sandbox execution environments
- Collaborate with product managers to translate research insights into product features and roadmap priorities
- Stay current on latest advancements in agentic AI research, specifically in large language models, multi-agent systems, chain of thought reasoning, and autonomous agents
Research Publication & Innovation
- Author technical papers for top-tier conferences on agent orchestration, context optimization, RAG-based embeddings, and real-time data integration
- File patents for novel techniques in agent-data interaction, token optimization, and CodeAct patterns
- Present research findings at internal tech talks and external conferences
- Mentor engineers and junior scientists on machine learning techniques, experimental design, and research methodologies
A day in the life
You start your morning analyzing experiment results from overnight runs comparing three evaluations for different RAG-based embedding approaches. The data shows that one of the embedding pattern is returning a significant improvement in accuracy

You create a spec file with the findings and start drafting a technical paper to be shared with Amazon AI forum.
Mid-morning, you're in a design session with the engineering team discussing how to optimize RAG-based embeddings for semantic search over advertiser data. You propose using a hybrid approach combining dense and sparse embeddings to represent campaign metadata, enabling agents to find relevant campaigns through natural language queries while maintaining sub-second latency. You sketch out the architecture and discuss trade-offs between embedding model size, search latency, and accuracy.
After lunch, you dive into advertiser interaction logs from advertising agents and skills

You're looking for patterns in how advertisers ask questions about their campaigns. You discover that 60% of queries follow a similar structure: filter campaigns by criteria, aggregate metrics, and compare to benchmarks. This insight leads you to design a new pre-computation strategy using RAG-based embeddings that could reduce query latency by 40%.
In the afternoon, you collaborate with an Applied Scientist from an advertising agent team

They're seeing inconsistent results when agents try to calculate complex metrics across multiple campaigns. You investigate and discover the issue is related to how the agent interprets the advertiser context. You propose enriching the RAG-based embeddings with richer metadata descriptions and run experiments showing this improves calculation accuracy from 85% to 98%.
Late afternoon, you're prototyping a new approach for adaptive context selection using RAG-based embeddings with the spec file you generated earlier

Instead of providing agents with all available advertiser data, you want to dynamically select the most relevant datasets based on query intent using semantic similarity. You build a quick proof-of-concept and test it on historical queries. The results are promising: 30% reduction in tokens with no loss in response quality.
About the team
The Ads Real-Time Data Service team is a diverse group of passionate engineers and scientists dedicated to advancing agent-data interaction technology for advertising AI

We value creativity, collaboration, and a commitment to excellence. Our team thrives on tackling complex problems at the intersection of real-time data engineering, AI agent systems, and large language model optimization-turning innovative research ideas into production systems that serve millions of advertisers.
We are highly motivated, collaborative and fun-loving with an entrepreneurial spirit and bias for action. We have a broad mandate to experiment and innovate, working on problems in agentic AI, context optimization, RAG-based embeddings, and real-time data delivery

We celebrate both research excellence (papers, patents) and engineering impact (production systems serving 30+ advertising agents and skills). We maintain a sustainable pace with flexible work arrangements and a strong focus on work-life balance.