Responsibilities
- Prototype, iterate, and productionize reusable agent modules for defined enterprise workflows.
- Maintain versioned prompts, policies, tool schemas, configurations, change logs, and reproducible agent assets.
- Optimize agent latency and token efficiency, including for edge-targeted scenarios where applicable.
- Build regression suites, golden test sets, comparison reports, evaluation metadata, and quality and safety checks.
- Implement and extend MCP clients and connectors for internal data products and approved enterprise applications.
- Validate integrations using sandbox credentials, representative test data, and end-to-end workflow tests.
- Ensure components meet operational readiness expectations through logging, telemetry, runbooks, and basic SLI/SLO alignment.
- Collaborate with platform, transformation, product, domain, security, SRE/observability, and enterprise application teams.
- Improve repeatability through evaluation templates, prompt-versioning conventions, and integration-test scaffolds.
Requirements
- Bachelor’s or Master’s degree in Computer Science, AI, ML, or Data Science, or equivalent experience.
- Hands-on experience building LLM applications involving agents, tool use, and prompting.
- Production Python engineering experience with testing, packaging, and structured logging.
- Experience with agentic AI frameworks and patterns such as LangGraph, LangChain, CrewAI-style orchestration, tool/function calling, and prompt versioning.
- Experience with test sets, regression testing, offline evaluation metrics, A/B comparisons, and failure taxonomies.
- Experience with APIs, authentication concepts, schema-based tool integration, and preferably MCP-style interface implementation.
- Understanding of correlation IDs, error analysis, latency instrumentation, and cloud deployment and validation through platform pipelines.
- Ability to take components from prototype through testing to production readiness and improve repeatability through templates and automation.
- Ability to collaborate with domain teams, adapt to changing model and tool constraints, and explain agent behavior and evaluation results to non-experts.
Tech Stack
Categories
About PepsiCo
PepsiCo is a playground for curious people. We invite thinkers, doers, and changemakers to champion innovation, take calculated risks, and challenge the status quo. From executives to team members on the front lines, we’re excited about the future. We take chances. Together, we dare to make the world a better place. Our associates are the magic ingredient. Each of them plays an integral role in helping create deep connections between people and our products. Think about your last group celebration: Chances are, one of our iconic brands was by your side. At PepsiCo, you’re invited to be a part of a global team of innovators who make, move, and sell these products—which are enjoyed by more than 1 billion people a day. A career at PepsiCo means working in a culture where everyone’s welcome. Here, you can dare to be yourself. No matter who you are or where you’re from, you can influence the people around you and the world at large. By showing up, you’ll have the opportunity to learn, develop and grow your skills for the future. Our supportive teams can fuel your professional goals to make a global impact on people and the planet. Join us. Dare for Better.