Role Overview
As a Staff Machine Learning Engineer at MediaAlpha, you will lead the design and development of LLM-powered applications for our internal teams and ad serving platform. You will be accountable for end-to-end systems: owning strategy, architecture, implementation, and operations. You will collaborate with engineering, product, and business stakeholders to deliver intelligent systems that solve real business problems.
The ideal candidate is a strong engineer with hands-on experience building LLM-powered applications in production, who works well across organizational boundaries and improves the conditions of any area they are in. We are an intentionally small organization, designed for agility, speed, and workability.
Responsibilities
- Own the full software development lifecycle of LLM-powered systems, from requirements through production deployment and operational health.
- Design and implement retrieval-augmented generation (RAG) systems, tool-use patterns, and multi-step reasoning workflows.
- Integrate LLM-based services into MediaAlpha's products, internal tools, and data ecosystem.
- Drive architectural decisions balancing capability, cost, latency, and maintainability.
- Establish evaluation frameworks and metrics to measure application quality and business impact.
- Work directly with stakeholders to define product capabilities and develop new features and integrations.
- Be a champion of instilling excellence in coding, testing, and quality into the team and culture.
- Mentor engineers on LLM application development patterns and best practices.
Requirements
- 9+ years of experience in software development.
- Experience designing and owning systems end-to-end, including 3-tier architectures, data flows, APIs, deployments, and operational health.
- Strong programming skills in Python, including software engineering methodologies and best practices.
- Hands-on experience building LLM-powered features in production.
- Experience with RAG architectures, including vector databases, embedding models, and retrieval strategies.
- Practical prompt engineering skills with demonstrated ability to iterate and evaluate quality.
- Familiarity with core ML concepts (classification, evaluation metrics, embeddings) and the judgment to know when traditional ML approaches outperform LLM-based solutions.
- Experience with agent frameworks and tool-use patterns (MCP, function calling, etc.).
- Familiarity with cloud environments (AWS preferred: EC2, S3, Bedrock).
- A self-starter who works independently, is comfortable with ambiguity, and pays careful attention to detail.
- Experience collaborating with engineers and non-technical stakeholders.
- Ability to articulate and communicate clearly, concisely, and thoughtfully.
- Organized and self-motivated.
- B.S. required, preferably in a technical discipline (e.g., computer science).