The posting
AI Engineer
Mid(4-7 years)
Senior(8-12 years)
Role Overview
Support the design, testing, and optimization of prompts and GenAI-driven workflows. Focus on improving output accuracy, structured responses, and supporting document extraction and business automation use cases.
Lead the design, governance, and optimization of enterprise-grade Gen AI and agentic workflows. Architect multi-agent systems that drive measurable automation outcomes, while ensuring reliability, scalability, and responsible AI practices.
Key Responsibilities
- Design and refine prompts for LLM-based applications
- Create structured output formats (JSON, schema-based responses)
- Test and evaluate AI outputs for accuracy and consistency
- Support development of simple multi-step AI workflows
- Assist in building reusable prompt libraries
- Support RAG-based document extraction workflows
- Document AI behaviour patterns and improvement cycles
- Work closely with business teams to refine outputs
Prompt & AI Workflow Architecture
- Design advanced prompt engineering frameworks
- Design and develop prompt chaining and tool-calling workflows
- Develop structured output enforcement frameworks
- Optimize AI behavior for enterprise use cases
Agentic Orchestration
- Design and implement multi-agent orchestration frameworks.
- Design and implement agentic workflows and orchestration layers
- Build AI orchestration services integrating multiple agents
- Develop and manage:
- RAG pipelines
- Embedding workflows
- Real-time document processing pipelines Define Agent Skills (task-specific capabilities assigned to agents)
- Implement Agent-to-Agent (A2A) communication patterns
- Design coordination logic across multiple agents
- Define escalation and fallback mechanisms
Enterprise AI Integration
- Work with engineering teams to integrate:
- Agent registries
- MCP connectors (Model Context Protocol or contextual connectors)
- Knowledge sources and enterprise systems
- Design contextual memory strategies
Optimization & Governance
- Implement hallucination control and guardrails
- Develop AI evaluation and benchmarking frameworks
- Define AI best practices and governance standards
- Improve automation KPIs (e.g., STP improvement)
- Ensure explainability and auditability
Evaluation & Optimization
- Define and implement AI evaluation frameworks (accuracy, consistency, completeness)
- Establish feedback loops using user inputs and production data
- Identify underperforming prompts and drive improvement cycles
- Improve automation KPIs such as Straight-Through Processing
- Analyze trends (e.g., prompt accuracy degradation, edge cases) and recommend fixes
Leadership
- Mentor junior AI engineers
- Collaborate directly with business stakeholders
- Translate business requirements into AI agent designs
- Drive AI adoption maturity across functions
Technical Requirements
- Strong understanding of LLM fundamentals
- Experience with:
- Few-shot prompting
- Structured prompting
- Prompt chaining (basic level)
- Basic Python knowledge
- Familiarity with:
- Azure OpenAI or similar LLM platforms
- Vector search concepts
- Exposure to:
- AI output evaluation methods
- Business process automation
AI & LLM Expertise
- Deep understanding of LLM behavior and limitations
- Strong expertise in:
- Prompt chaining
- Tool invocation frameworks
- Structured response validation
- Context window optimization
Agentic Systems
- Hands-on experience with:
- Agent orchestration frameworks
- Multi-agent collaboration models
- Agent Skills modeling
- Agent registries
- Agent-to-Agent (A2A) integration
- Familiarity with:
- MCP connectors or contextual integration frameworks
- Cross-system context injection
Azure Ecosystem
- Strong familiarity with Azure AI stack:
- Azure OpenAI
- Azure Cognitive Search
- Azure AI Foundry
- Understanding of how AI integrates into:
- Azure Functions
- API-based workflows
Enterprise Architecture
- Experience implementing RAG architecture
- Experience with embeddings & vector databases
- Strong understanding of Azure AI ecosystem
- Python proficiency for experimentation and prototyping
Preferred Qualifications
- Exposure to document processing
- Basic understanding of multi-agent systems
- Familiarity with AI safety and hallucination control concepts
- Strong written communication skills
- Optional: Insurance domain knowledge (Claims/UW)
- Experience deploying agentic AI in production
- Exposure to Responsible AI frameworks
- Strong stakeholder management skills
- Experience defining AI governance operating models
- Optional: Insurance domain expertise (Claims, UW workbench)
