Data Platform Architect (AgentExchange)

Join Salesforce as a Data Platform Architect for AgentExchange, where you'll consolidate fragmented data pipelines into a trusted data platform. You'll engage with executive stakeholders, set the long-term technical direction, and lead cross-org technical initiatives. Your success will be measured by the unification of analytics destinations, the production of an agentic data layer, and the implementation of predictive models.

• Consolidate fragmented pipelines and destinations into one trusted data platform, defining contracts for every team.
• Architect the LLM- and agent-native layer that turns marketplace data into intelligent experiences for partners, customers, and Salesforce leadership.
• Engage with executive stakeholders, represent the data platform in cross-org architecture reviews, and set the long-term technical direction for the data platform.
• LLM systems experience, in production: RAG, embeddings and vector stores, prompt and context engineering, offline and online evaluation, cost and latency tuning, hallucination and safety controls
• 12+ years in software / data engineering, including multi-year ownership of an enterprise-scale data or ML platform
• Deep architecture experience in at least three of: lakehouse / warehouse design, streaming + batch pipelines, dimensional and event modeling, feature stores, model serving
• Cloud-native data infrastructure: Snowflake, BigQuery, Redshift, or Databricks; AWS-based platforms
• Data security and governance as a first-class skill: PII classification, multi-tenant isolation, fine-grained access control, GDPR / CCPA, lineage and audit, and the security implications of LLM / agent access patterns
• NPS and effort-score measurement architecture at scale
• Privacy-preserving ML (differential privacy, tokenization, synthetic data)
• Familiarity with Salesforce Platform features and best practices
• A related technical degree required
• Salesforce Data 360, Tableau Next, Slack, MuleSoft data integration
• Agent evaluation frameworks and LLM observability (traces, eval datasets, regression suites)
• Track record representing a technical domain in cross-org architecture forums and influencing direction across teams you don't manage
• Working knowledge of MCP or equivalent tool / agent protocols, and a clear point of view on exposing data to agents safely
• Preferred
• Executive communication: can defend an architecture to a CTO and explain trade-offs to a PM in the same hour
• Marketplace or e-commerce data: GMV, attrition, conversion funnels, search signal processing
• Large-scale migrations (Heroku → cloud-native) with zero production disruption
• In office expectations are 10 days/a quarter to support customers and/or collaborate with their teams