Senior AI Engineer, Agentic AI for Finance
Pure Storage5 hours ago
Bengaluru, IndiaSenior
Responsibilities
- Architect, deploy, and own production-grade multi-agent workflows using LangGraph and Amazon Bedrock AgentCore.
- Build secure least-privilege Model Context Protocol tools and enterprise integrations across ERP, CRM, and ticketing systems.
- Develop grounded retrieval architectures and read-only Snowflake query layers for trusted operational insights.
- Create performant Next.js portals, administrative consoles, and human-in-the-loop review interfaces.
- Implement automated evaluations, shadow-mode validation, role-based access controls, audit logging, and prompt-injection defenses.
- Monitor accuracy, drift, latency, and cost with Datadog and Langfuse, while guiding engineers and vendors through reviews and architecture work.
Requirements
- Deep hands-on experience designing, shipping, and operating complex multi-agent architectures and LLM applications in production.
- Mastery of Python, TypeScript, SQL, LangGraph, and AWS services including Bedrock, AgentCore, Lambda, and ECS.
- Experience developing secure MCP tools and APIs, grounded retrieval/RAG pipelines, and fine-grained authorization for enterprise data.
- Ability to implement security controls, audit logging, human-in-the-loop verification, and continuous automated evaluation frameworks.
- Ability to translate financial and business processes into system specifications, assess technical trade-offs, and guide contributors through architecture reviews and mentorship.
Benefits
- Primarily in-office work from the Bangalore office, except during PTO, work travel, or approved leave.
- Flexible time off, wellness resources, and company-sponsored team events.
- Growth support, an inclusive workplace, employee resource groups, and accessibility accommodations.
Tech Stack
Categories
About Pure Storage
Pure Storage builds all-flash data storage systems and software for enterprises, including FlashArray and FlashBlade, and offers storage delivered as-a-service for on‑prem and cloud workloads. Its subscription model (Evergreen/Pure as‑a‑Service) supports databases, virtualization, analytics, and AI. Founded in 2009 and headquartered in Santa Clara, California, Pure is a public company listed on the NYSE (PSTG).