Procter & Gamble is seeking a Senior Data Architect to join the Supply Chain Innovation organization. This role focuses on leading the design of enterprise-scale Generative AI solutions integrated with P&G's data platform. You will own the architecture for how AI systems retrieve, reason over, and act on enterprise data, setting standards for engineering teams.
Key Responsibilities:
Own end-to-end architecture for enterprise GenAI solutions, including retrieval, orchestration, integration, and evaluation.
Design production RAG architectures, including document ingestion, chunking, indexing, hybrid retrieval, and relevance tuning.
Architect multi-phase agent and tool-calling workflows with deterministic guardrails.
Design data ingestion and synchronization foundations for AI grounding, including streaming and change data capture.
Enforce data governance, multi-tenant isolation, and Responsible AI principles.
Establish evaluation and observability standards for AI workloads, including automated quality scoring and production tracing.
Manage operational economics, including latency, throughput, and inference costs.
Required Qualifications
Bachelor's degree in Computer Science or related field.
10+ years of experience in software engineering with a focus on architecture and technical leadership of large-scale, high-throughput platforms.
Hands-on experience delivering production Generative AI solutions (RAG, LLM-based classification, agent workflows).
Experience architecting orchestration layers for distributed systems.
Strong background in search and retrieval (vector/semantic similarity, hybrid retrieval).
Experience defining quality and evaluation practices for probabilistic systems.
Proficiency in at least one backend language (Python, Node.js/TypeScript).
Cloud-native engineering experience (Azure) with containerization and CI/CD.
Preferred Qualifications
Knowledge of modern lakehouse platforms (Databricks, Delta Lake, Unity Catalog).
Familiarity with AI governance and designing for regulated or highly sensitive data environments.
For full details about this role, click “Apply on company site” to view the employer’s original posting.
A focused preparation guide, built around this job’s responsibilities and requirements.
✦ AI-generated guide
Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.
01
Understand the role
What matters most
Architecting Reliable GenAI Workflows
The role requires designing end-to-end RAG and agentic systems that move beyond experimentation into production. You must prioritize building deterministic guardrails, structured output schemas, and robust fallback logic to ensure non-deterministic LLMs meet enterprise reliability standards.
Data Foundation and Grounding
You are responsible for the data ingestion and synchronization architecture that powers AI grounding. This involves designing streaming pipelines, change data capture, and maintaining canonical data models to ensure AI systems reason over current, accurate enterprise data.
Governance and Multi-Tenant Isolation
Given P&G's global scale, you must enforce strict data governance and multi-tenant isolation. This includes implementing access controls, PII protection, and Responsible AI principles to ensure secure data handling across diverse brands and regions.
Operational Economics and Observability
You must define standards for AI performance, including latency, throughput, and inference cost management. Establishing automated evaluation, model-based judging, and production tracing is critical to maintaining system quality and cost-efficiency.
Likely challenges to think through
Balancing Determinism with Probabilistic Systems
A core challenge is applying traditional distributed systems engineering—such as idempotency and state management—to probabilistic AI models. Prepare by demonstrating how you have successfully integrated rigid software engineering practices with the inherent variability of LLM outputs.
Managing Complex Data Freshness at Scale
Maintaining real-time or near-real-time grounding for AI systems across heterogeneous enterprise sources is difficult. Prepare to discuss strategies for handling data drift, incremental indexing, and the trade-offs between retrieval latency and data freshness.
Influencing Without Formal Authority
As a principal technical authority, you will set standards for other engineering teams. Prepare to explain how you drive consensus on architectural decisions and technical debt management across global, cross-functional teams without having direct management control over all stakeholders.
03
Practice your answers
Technical practice
Designing Production RAG Architectures
What to coverExplain your approach to balancing lexical and semantic retrieval in a hybrid system. Discuss how you handle document chunking strategies and re-ranking to ensure high relevance in enterprise-scale datasets.
Deterministic Guardrails for Agent Workflows
What to coverHow do you implement deterministic guardrails within non-deterministic LLM agent workflows? Focus on techniques like structured output schemas, validation logic, and fallback mechanisms to ensure production reliability.
Data Grounding and Synchronization
What to coverDescribe your strategy for maintaining data freshness in AI systems using streaming and change data capture (CDC). Address how you ensure consistency between enterprise source systems and AI indexing layers.
Evaluation and Observability for Probabilistic Systems
What to coverWhat framework would you establish to evaluate LLM performance in production? Discuss the use of curated test sets, model-based judging, and end-to-end tracing to monitor system behavior and drift.
Managing Operational Economics
What to coverHow do you optimize for latency, throughput, and inference costs in a high-scale AI architecture? Suggest strategies for caching, model routing, and cost-efficient infrastructure design.
Multi-tenant Isolation and Governance
What to coverHow do you enforce strict data isolation and governance in a shared AI platform? Discuss your approach to PII protection and ensuring compliance with Responsible AI principles across different business units.
Distributed Systems Orchestration
What to coverGiven your experience with distributed systems, how do you handle partial failures and idempotency in complex agent workflows? Explain how you apply traditional distributed systems patterns to AI orchestration layers.
Cloud-Native AI Engineering
What to coverDescribe how you integrate CI/CD pipelines with AI model deployment on Azure. Focus on how you manage containerization and automated regression testing for probabilistic components.
Behavioural practice
Influencing Stakeholders Without Formal Authority
What to coverDescribe a situation where you had to align cross-functional teams on a complex technical architecture. How did you translate business needs into technical requirements and gain consensus?
Handling Production Failures in Distributed Systems
What to coverProvide an example of a time you had to troubleshoot a critical failure in a high-throughput system. What was the situation, the specific actions you took to resolve it, and the long-term architectural changes you implemented?
Mentoring and Technical Leadership
What to coverTell me about a time you mentored engineers to adopt new standards or technologies. How did you balance hands-on development with your responsibilities as a technical lead?
Navigating Ambiguity in AI Projects
What to coverDescribe a project where you had to design a solution for a non-deterministic problem. How did you define success metrics and manage stakeholder expectations throughout the development lifecycle?
02
Sharpen your knowledge
Topics to prepare
Production RAG Architecture
Focus on the end-to-end pipeline including document ingestion, chunking strategies, hybrid retrieval (lexical and semantic), and re-ranking. Exercise: Design a system diagram for a RAG pipeline that handles incremental data updates via change data capture (CDC) and explain how you would maintain index freshness.
AI Observability and Evaluation
Understand how to measure the performance of probabilistic systems using curated test sets, model-based judging, and production tracing. Exercise: Define a set of metrics and a testing framework to evaluate the quality of LLM-generated responses in a multi-tenant environment.
Agentic Workflow Reliability
Study deterministic guardrails for agentic systems, including tool-calling, structured output validation, and fallback mechanisms. Exercise: Draft a state machine or workflow logic for an agent that requires human-in-the-loop intervention for specific high-risk actions.
Skills to revise
Azure Cloud-Native Engineering
Demonstrate proficiency in deploying scalable AI workloads on Azure, including containerization and CI/CD integration. Exercise: Outline a CI/CD pipeline strategy that incorporates automated regression testing for LLM prompts and model outputs.
Data Governance and Multi-tenancy
Apply principles of data isolation and security in enterprise AI systems. Exercise: Propose an architecture for enforcing strict access control and PII protection across different business units sharing a centralized data lakehouse.
Lakehouse Data Modeling
Gain familiarity with Databricks, Delta Lake, and Unity Catalog for managing enterprise data. Exercise: Explain how to structure a medallion architecture (bronze, silver, gold) to support efficient retrieval for GenAI applications.
04
Your preparation checklist
Prepare a RAG Architecture Case Study
Develop a technical walkthrough of a production-grade RAG system you have designed. Focus on your approach to document chunking, hybrid retrieval strategies, and relevance tuning. Be prepared to explain how you handled data freshness and synchronization from enterprise sources.
Define Evaluation Frameworks for Probabilistic Systems
Outline your methodology for measuring the quality of LLM outputs. Prepare to discuss how you implement automated scoring, model-based judging, and regression testing within a CI/CD pipeline to ensure reliability in production environments.
Draft a Strategy for AI Guardrails and Governance
Prepare to discuss how you architect multi-tenant isolation and enforce data governance. Specifically, explain how you design deterministic guardrails for agent workflows to handle sensitive enterprise data while maintaining compliance and security.
Formulate Operational Economics and Scalability Plans
Be ready to explain how you optimize for latency, throughput, and inference costs. Prepare examples of how you have managed resource allocation in distributed systems and how you would apply these principles to manage AI model costs at scale.
Prepare Strategic Questions for the Interviewers
Draft thoughtful questions regarding P&G's current data platform maturity, such as how they leverage Databricks or Unity Catalog for AI grounding, and how the team balances rapid innovation with the strict governance requirements of a global consumer goods company.
Availability last checked: 18 Oct 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy
Share this opportunityKnow someone who would be a great fit?