Manager, Analytics & Agentic Platform Operations

Visa Visa · Fintech · San Francisco, CA +1

Manager role focused on building and operating an end-to-end data platform using Microsoft Fabric and Azure, with a strong emphasis on AI-driven infrastructure and agent orchestration. The role involves designing data pipelines, developing AI agents for platform maintenance and validation, and ensuring security, cost optimization, and observability for both data and AI workloads.

What you'd actually do

  1. Act as an agent orchestrator—leveraging AI to build out and maintain the platform itself. Design and run AI agents on automated loops, and develop self-running frameworks that continuously validate the environment and flag inconsistencies (security drift, schema changes, broken lineage) before they become problems. Build and operationalize MCP servers, automation agents, and copilots that enable natural-language data access and self-healing pipelines.
  2. Design, build, and orchestrate data pipelines and Spark/PySpark notebooks within Microsoft Fabric—ingesting, transforming, and moving data across the Bronze/Silver/Gold medallion architecture to deliver reliable, production-grade datasets.
  3. Implement and maintain granular security protocols to protect sensitive data. This includes writing and managing scripts for Row Level Security (RLS), Column Level Security (CLS), and Object Level Security (OLS) to ensure strict data isolation and PII protection across the platform.
  4. Take direct ownership of cloud and AI spend. Track daily consumption against budgets, identify “runaway” queries and inefficient agent/LLM token usage, and optimize capacity, model, and query settings to prevent billing overages.
  5. Implement monitoring alerts (using SQL or third-party tools) to detect “silent” data failures—such as stale data, schema drift, or volume anomalies—before they impact business users. Extend the same rigor to AI workloads—monitoring agent and model behavior for drift, errors, and runaway loops.

Skills

Required

  • Microsoft Fabric
  • Azure Data Stack
  • PySpark
  • SQL
  • Python
  • Data Engineering
  • DevOps
  • CI/CD
  • Security protocols (RLS, CLS, OLS)
  • Cloud cost management
  • LLMOps/MLOps observability

Nice to have

  • Hadoop
  • Bash/Shell scripting
  • SSH
  • Trino
  • Hive SQL
  • Airflow
  • PowerShell
  • Claude Code
  • Codex
  • LangSmith

What the JD emphasized

  • agent orchestrator
  • AI-driven infrastructure
  • agentic CLI coding tools
  • regulated environment

Other signals

  • agent orchestration
  • AI-driven infrastructure
  • self-operating future
  • LLM-powered agents
  • agentic CLI coding tools