Staff Engineer Software - AI/Agentic AI [T500-30275]

About Albertsons:

As a leading food and drug retailer in the United States, Albertsons Companies, Inc. (ACI) operates over 2,200 stores across 34 states and the District of Columbia. Our well-known global banners, including Albertsons, Safeway, Vons, Jewel-Osco and others. Our success is built upon a united team approach, driven by the desire to understand and enhance the customer experience. This customer-centric focus not only inspires innovation but has also led to remarkable achievements across our digital platforms and omnichannel strategies. By constantly pushing the boundaries of retail, we have transformed shopping into an experience that is not just easy and efficient, but also fun and engaging. Over the past year, we have taken bold, strategic steps to migrate and modernize our enterprise data hub on Google Cloud, revolutionizing our data capabilities, thus becoming the first fully cloud-based grocery tech company.

About Tech Hub in India:

Albertsons Companies India (ACI India) is part of Albertsons Companies Inc. global workforce. As Albertsons’ first significant strategic investment in India, it serves as a collaborative hub, supporting Albertsons in driving innovation and efficiency across Technology, Digital, and other key business functions. The employees at the India GCC will play a pivotal role in enhancing the company’s digital transformation, customer engagement, and operational excellence.

We are excited to announce an opening for Staff Software Engineer, AI Engineering (AIDLC) at ACI. Please find below the details of the role and its responsibilities.

Roles & responsibilities:

• Build and own production agents end to end for the AIDLC (AI-driven development lifecycle) pod: design, build, evals, release, monitoring and on-call.
• Build agents for stages of the software development lifecycle: requirements and design, code review before merge, security-finding remediation, test generation, release and path to production, and SRE and incident support.
• Design and build multi-agent workflows with clear task decomposition, handoffs, shared state and failure handling, following the architecture patterns set by the Senior Staff engineer.
• Build and use MCP servers and tools, including code knowledge graphs exposed over MCP, so agents can work on Java/Spring and React/Angular codebases.
• Implement memory and state management, context engineering and RAG for your agents.
• Write the evals for your agents: golden datasets, LLM-as-judge scoring and regression suites. Agent changes ship only when evals pass.
• Instrument your agents with tracing and logging, and use traces to find and fix failures in production.
• Implement guardrails: least-privilege tool access, prompt-injection defence and human-in-the-loop approval points.
• Review agent pull requests from SE I and SE II engineers and mentor them on agent design and debugging.
• Write production code every week. Use agentic coding tools daily (e.g. Claude Code, GitHub Copilot, Cursor).
• Report the effect of your agents against the pre-agent baseline: delivery speed and cost, defects, security findings and cycle time.
• Track token and compute cost for your agents and keep it within budget.
• Work with Security, Cloud, Data and Product partners so your agents meet security, risk and compliance standards.

Experience Required:

• AI experience is a MUST. You have built and deployed at least one LLM agent or agentic workflow in production, used by real users, with evals and tracing in place, and you owned it end to end. Describe it in your application or link a repo or demo.
• 9-12 years of software engineering experience.
• Hands-on programming in Python, Java or TypeScript, used to build agents and the services around them. The agents work on Java/Spring and React/Angular codebases.
• Multi-agent orchestration is required: delegation, handoffs, shared state and failure handling. Frameworks are examples only (e.g. LangGraph, LangChain/LangSmith, Claude Agent SDK, OpenAI Agents SDK, Google ADK).
• Hands-on experience with tool/function calling, MCP (building and using servers), memory and state management, context engineering and RAG.
• Experience writing evals: golden datasets, LLM-as-judge and regression testing for agents.
• Experience with tracing and observability for LLM applications, and with guardrails, prompt-injection defence and human-in-the-loop design.
• Everyday use of agentic coding tools (e.g. Claude Code, GitHub …