Responsibilities
- Design and build internal AI agents and automation workflows using technologies such as LangChain, LangGraph, Snowflake Cortex, LlamaIndex, or similar frameworks, to support planning, tool use, retrieval, validation, and human-in-the-loop execution where appropriate.
- Develop reusable tools, APIs, and components that engineers can compose into new agentic workflows.
- Implement retrieval-augmented generation workflows, context management strategies, and prompt patterns that improve accuracy, reliability, latency, and cost efficiency.
- Build evaluation harnesses, regression tests, and monitoring patterns for AI-agent behavior.
- Define and track metrics such as task completion, groundedness, response accuracy, latency, cost, and failure rate.
- Design guardrails and validation patterns to reduce hallucinations, unsafe outputs, and unreliable automation behavior.
- Partner with engineering and operations teams to move AI workflows from prototype to production-ready systems.
- Design, build, and maintain production-grade Python applications, libraries, and services.
- Champion object-oriented design principles, including encapsulation, abstraction, inheritance/composition, reusable interfaces, and clean separation of concerns to improve maintainability and extensibility.
- Champion software engineering best practices including modular design, automated testing, CI/CD, code reviews, observability, and documentation.
- Create reusable engineering patterns that reduce bespoke development effort and improve consistency across partner engagements.
- Collaborate with product, engineering, operations, and data platform teams to translate repeatable business needs into scalable technical solutions.
- Build reusable Python libraries that support clean room capabilities across first-party audience workflows and core measurement use cases, including audience onboarding, ingestion, indexing, activation, campaign and impression delivery analysis, reach and frequency, attribution, and incrementality.
- Abstract complex analytical and data collaboration workflows into repeatable, self-service components for internal teams and external partners.
- Enable configurable feature deployment so new audience and measurement capabilities can be delivered quickly and consistently across partners.
- Mentor engineers through code reviews, technical design discussions, and operational best practices.
- Help establish engineering standards for AI-assisted workflows, agentic system design, reusable libraries, and production automation.
- Promote a culture of reliability, maintainability, and continuous improvement.
Requirements
- Strong Python engineer
- Ability to design reliable systems
- Experience in productizing repeatable workflows
- Experience building AI agents that help automate complex engineering and operational tasks across clean rooms for audience activation, measurement, and reporting
