We are looking for an AI Specialist to design, build, and maintain AI agents that support our business teams across a retail network of 500+ stores. In this role, you will develop LLM-based agentic applications end-to-end, from understanding business workflows to deploying and maintaining reliable, secure, production-grade systems, and help establish the platform and standards behind them.
Key Responsibilities
1. AI Agent Design & Development
Design and develop AI agents using large language models (LLMs), including retrieval-augmented generation (RAG), tool/function calling, and structured outputs.
Build multi-agent workflows with clear hand-offs and shared context where appropriate.
Apply prompt engineering and grounding techniques to balance accuracy, reliability, and business relevance.
2. Platform & Reusable Components
Build reusable platform components — connectors, retrieval layers, tool libraries, and evaluation frameworks — to accelerate the development of future agents.
Maintain trusted, well-governed data layers that agents depend on.
3. API Development & System Integration
Develop high-performance Python APIs (FastAPI / Flask) to integrate AI capabilities into existing web, mobile, and enterprise systems.
Integrate with enterprise data sources and tools, including standardised protocols such as MCP (Model Context Protocol).
Implement concurrency and asynchronous processing to handle requests reliably at scale.
4. Cloud Infrastructure & Model Operations
Architect and maintain scalable solutions on AWS (e.g., Bedrock, SageMaker, Lambda, S3) and integrate model providers such as Anthropic and OpenAI.
Set up and manage vector / retrieval databases for RAG.
Manage cost-efficiency and performance across large-scale operations.
5. LLMOps, Evaluation & Lifecycle Management
Build evaluation pipelines covering task success, accuracy, and safety/hallucination checks.
Implement observability and tracing to monitor live agent behaviour, latency, cost, and quality.
Maintain reproducibility through Git, CI/CD, and Docker, and ensure continuous post-deployment improvement.
6. Security & Data Governance
Apply DevSecOps best practices, including security scanning, identity & access management (IAM), and data encryption.
Design for PII protection and data governance with auditable, least-privilege access.
Collaborate with IT and Security teams to ensure compliance with privacy and responsible-AI standards.
7. Business Impact & Maintenance
Partner with business stakeholders to translate operational challenges into well-scoped AI solutions and support their adoption.
Provide end-to-end maintenance for deployed applications, ensuring high availability and prompt resolution of production issues.
Research and evaluate emerging AI and agentic technologies to identify opportunities for measurable business impact.