Senior AI Engineer

Job Overview

We are looking for a highly capable Senior AI Engineer to lead the design and implementation of an AI layer for a modern Retail Execution Platform. This role will be responsible for building the intelligence foundation that enables conversational experiences, workflow assistance, contextual recommendations, and safe execution of business actions across the platform.

The ideal candidate will bring strong experience in LLM-based application design, MCPs, RAG, AI orchestration, enterprise-grade conversational systems, and workflow-aware AI integration. This role requires someone who can bridge modern AI capabilities with structured enterprise application behavior, ensuring that AI interactions are reliable, secure, tenant-aware, role-aware, and deeply aligned with platform business rules.

Key Responsibilities

AI Copilot Architecture and Orchestration

• Lead the architecture and implementation of the AI layer for the Retail Execution Platform.
• Define how large language models are orchestrated to support structured business interactions, workflow assistance, and action triggering.
• Design a scalable AI orchestration model that separates prompt intelligence, workflow routing, policy enforcement, and backend execution.
• Ensure the AI layer is enterprise-ready, maintainable, auditable, and aligned with platform architecture principles.
• Build the foundation for AI capabilities that can evolve over time without compromising control, trust, or operational quality.

Conversational Workflow Design

• Define how natural language interactions are translated into business workflows and executable platform actions.
• Establish patterns for converting conversational intent into structured commands that can be safely interpreted by backend services.
• Design interaction models that support user guidance, confirmations, approvals, and progressive execution of tasks.
• Ensure conversational flows are predictable, explainable, and aligned with business process expectations.
• Create AI interaction patterns that support action-taking without losing transparency or user control.

Intent-to-Command and Backend Integration

• Design structured contracts between AI-generated intent and platform service execution.
• Work closely with backend teams to ensure AI commands can be safely routed to CAP-based business services.
• Define schemas and structured payload expectations that allow conversational requests to become deterministic backend operations.
• Ensure the AI layer communicates with application services through governed, typed, and secure interfaces.
• Build robust AI-service integration patterns that reduce ambiguity and improve execution reliability.

Context-Aware AI Experience

• Define and implement the context model that powers AI responses and decisions across the platform.
• Ensure AI behavior is informed by relevant business and runtime context such as tenant, user role, schedule, visit status, targets, and workflow stage.
• Design context-aware AI interactions that feel useful, relevant, and operationally grounded.
• Prevent generic or context-blind AI behavior by structuring how business signals are exposed to the model.
• Ensure contextual awareness improves decision quality without compromising data boundaries or security rules.

Prompt Engineering Standards and AI Guardrails

• Establish prompt engineering patterns, standards, and reusable templates for platform AI use cases.
• Define AI guardrails that reduce unsafe outputs, unsupported actions, and inconsistent behavior.
• Create strong controls around prompt quality, system instructions, fallback handling, and output structure.
• Ensure prompts are designed for enterprise reliability, not just conversational fluency.
• Continuously improve prompt performance based on production feedback, observed usage, and operational quality trends.

Safety, Approval, and Execution Control

• Define and implement safety boundaries for AI-assisted and AI-triggered actions.
• Establish clear approval and confirmation flows for actions that should not be executed automatically.
• Ensure AI execution remains bounded by business rules, policy controls, and user permissions.
• Build mechanisms for safe escalation, clarification, fallback, and human confirmation where needed.
• Design the Copilot experience to be useful and proactive without becoming uncontrolled or opaque.

Authentication, Tenant Isolation, and Authorization-Aware AI

• Integrate AI orchestration with enterprise authentication and tenant-aware runtime context.
• Ensure all AI queries, responses, and actions are resolved strictly within tenant boundaries.
• Implement authorization-aware filtering so the AI only suggests or executes actions a user is permitted to access.
• Prevent data leakage, cross-tenant contamination, and unauthorized action execution through strong AI-layer controls.
• Maintain a secure and compliant AI execution model suitable for enterprise SaaS environments.

Structured Output and Deterministic AI Behavior

• Define structured JSON-based interaction contracts for AI outputs that feed application workflows and UI behavior.
• Ensure AI-generated outputs are machine-readable, validated, and actionable where deterministic execution is required.
• Reduce free-form ambiguity in operational use cases by enforcing output structure and downstream validation.
• Build systems that distinguish clearly between informative AI responses and executable AI instructions.
• Improve reliability by combining LLM flexibility with strong application-level structure.

Canonical Data Alignment and Read Model Design

• Ensure the AI layer operates on normalized, platform-governed business data rather than raw upstream system structures.
• Align AI behavior with the Canon Core model and other platform-standard business entities.
• Define AI-friendly read models for operational domains such as visits, customers, schedules, targets, and related execution context.
• Implement abstraction layers that allow AI prompts and tools to access trusted business information in a consistent form.
• Protect the AI experience from source-system complexity by exposing only curated and normalized platform views.

AI Query Abstraction and Business Data Access

• Design data access patterns that allow AI components to retrieve the right business context safely and efficiently.
• Implement abstraction layers that prepare business data for prompting, reasoning, summarization, and action recommendation.
• Ensure AI interactions are powered by governed platform data rather than raw integration payloads.
• Define how source-of-truth principles are represented in AI responses, especially in cases of data conflict or reconciliation.
• Ensure AI messaging remains accurate and transparent when business state depends on synchronized enterprise systems.

Copilot Experience in the Application

• Support integration of the Copilot experience into the end-user application environment.
• Define how AI interactions are embedded into workflow-heavy application journeys in a natural and useful manner.
• Enable conversational support for planning, scheduling, task execution, order intent capture, lookup, progress guidance, and operational assistance.
• Design contextual confirmations and execution prompts that guide users through important decisions safely.

Required Experience

• 5+ years of software engineering experience, with strong recent experience in AI/LLM application development.
• Strong proficiency in Python for building AI/LLM-powered backend services.
• Strong understanding of Python frameworks for building APIs and service layers, such as FastAPI or similar backend frameworks.
• Proven experience designing and implementing MCPs, LLM orchestration and conversational application workflows.
• Strong experience building AI copilots, assistants, or workflow-integrated GenAI systems for enterprise or SaaS products.
• Experience defining structured output contracts, tool-calling patterns, or intent-to-action pipelines.
• Strong understanding of prompt engineering, guardrails, grounding, fallback behavior, and hallucination mitigation.
• Experience integrating AI systems with business services, APIs, and application workflows.
• Strong understanding of authorization-aware systems, tenant isolation, and enterprise security considerations.
• Experience with observability, audit logging, performance monitoring, and production support for AI-powered systems.