How will you make an impact?
You will own and build the core components of NICE’s AI platform, the integration layer, agent platform, Models Gateway, RAG pipelines, and developer tooling, working hands-on across the stack with the Architect, DevOps, and Security teams.
Build the MCP Integration Layer
• Implement MCP server and client libraries that connect enterprise systems (Atlassian, Microsoft 365, ServiceNow, Workday, Salesforce, Snowflake) to AI agents
• Design and expose clean tool schemas; handle auth flows (OAuth2, managed identity); implement error handling, retries, and rate limiting
• Build the A2A (Agent-to-Agent) interoperability layer enabling multi-agent collaboration across the platform
Develop the AI Agentic Platform
• Implement production-grade AI agent frameworks: ReAct loops, tool-augmented reasoning, multi-agent orchestration, memory and state management
• Build agent harnesses for specific NICE use cases: IT helpdesk automation, procurement workflows, HR self-service, developer productivity agents
• Integrate with Azure AI Foundry and Anthropic Claude API, managing context windows, tool use, streaming responses, and multi-turn conversations
Engineer the Models Gateway
• Build a unified gateway abstracting multiple LLM providers (Azure OpenAI, Anthropic, open-source models via Azure ML)
• Implement model routing logic, fallback chains, cost-based dispatch, latency budgeting, and per-team quota enforcement
• Add logging, token metering, and usage dashboards for FinOps visibility
Build RAG Pipelines & Vector Infrastructure
• Design and implement document ingestion pipelines: chunking, embedding generation, metadata enrichment, and upsert into vector stores
• Build retrieval pipelines with hybrid search (dense + sparse), re-ranking, and context assembly for LLM prompts
• Manage vector DB infrastructure on Azure AI Search and/other; own schema design and index optimization
Implement Prompt Management & LLM Evals
• Build a prompt registry: version control, templating engine, environment promotion, and rollback
• Design and run LLM evaluation pipelines: automated regression tests, hallucination detection, task-specific benchmarks
• Implement human-in-the-loop feedback collection and model performance tracking dashboards
Contribute to Developer Tooling & CI/CD
• Build and maintain GitHub Actions workflows for AI component testing, deployment, and rollback
• Write reusable SDK / client libraries for internal teams consuming the AI platform
• Integrate GitHub Copilot and Azure AI Foundry into the development workflow; document patterns for the broader R&D org
Observability & Production Operations
• Instrument all AI components with OpenTelemetry: traces, metrics, and structured logs
• Build Azure Monitor dashboards and alerts covering inference latency, error rates, token spend, and agent success rates
• Participate in on-call rotation for critical AI platform services