Responsibilities
- Build the data foundation for agentic solutions, including lakehouse architectures, metadata and semantic context, and the access controls that let custom agents act on trusted data.
- Stand up working proofs with customers’ data, including de-identifying sensitive data (e.g., PCI, PHI).
- Build and ship production data and machine learning workloads with customers, writing, testing, and performance-tuning PySpark, SQL, and Python on BigQuery, Managed Spark, Spanner, and AlloyDB.
- Train, evaluate, and deploy machine learning and agentic components where they improve the customer’s business metric, with evaluation and monitoring in place.
- Refactor and migrate high-cost analytics, Spark, and operational database workloads from other platforms onto Google Cloud engines, with measured price-performance results.
Minimum qualifications:
- Bachelor’s degree in Engineering, Computer Science, a related field, or equivalent practical experience.
- 8 years of experience with software development using Python or similar coding languages.
- Experience with machine learning or machine learning algorithms.
- Experience in data analysis (e.g., SQL, NoSQL, Python).
- Experience in data management and warehousing concepts or technologies (e.g., BigQuery).
Preferred qualifications:
- Experience in a consulting, client service, or customer-facing role.
- Experience with Google Cloud data products (e.g., BigQuery, Managed Service for Apache Spark, Dataflow, Spanner, AlloyDB, Dataplex, Looker) or equivalent platforms.
- Experience migrating analytics, Spark or database workloads between platforms with measured cost or performance outcomes.
- Experience implementing multi-agent systems using frameworks (e.g., LangGraph, CrewAI, ADK) and patterns (e.g., ReAct, self-reflection, hierarchical delegation).
- Experience in prompt development, model evaluation, and the application of AI.
- Knowledge of LLM-native metrics (e.g., tokens/sec, cost-per-request) and techniques for optimizing state management and granular tracing.