brAInspark is a live agentic AI platform that enterprises deploy multi-agent systems on. What you build runs inside their businesses, against their data, under their compliance regimes. You decide what the platform may do autonomously — and then prove it is safe to do it. THE ROLE You lead the AI engineering team behind Incedo’s agentic and GenAI products. You own the AI technology roadmap, the reference architecture for agentic systems across the portfolio, and the engineering standards the team builds to. This is a hands-on role. You will be in the code, in the evals and in the architecture reviews — not only in the roadmap deck.

1. Agentic and GenAI system design • Build production RAG, agentic and LLM-based systems — multi-agent orchestration, tool use, memory and MCP-based enterprise integration. • Define the agent reference architecture across Incedo’s platforms, so teams build on shared patterns rather than one per product. • Evaluate and integrate foundation models — what to call, what to fine-tune, and what should never be a model at all. 2. Evaluation, quality and governance • Own the evaluation strategy: offline evals, regression suites, human review loops and production LLM observability. • Establish guardrails — hallucination mitigation, PII-safe prompting, cost and latency budgets, and graceful fallbacks when confidence drops. • Set AI engineering standards for prompt versioning, reproducibility, testing and code quality. 3. Platform and delivery • Take solutions from prototype to production on cloud-native infrastructure. • Partner with Product, Solution Architecture and Engineering to turn roadmap intent into shippable increments. 4. Team leadership • Lead, mentor and grow the AI engineering team; run technical reviews and raise the technical bar. • Contribute to hiring and capability building as the team scales.

8+ years in software engineering, including 4+ years building AI/ML systems — and still hands-on. • Strong programming depth (Python preferred) with production-grade, reviewable code. • LLM and GenAI systems shipped to production, not only prototyped. • Agent frameworks in practice: LangGraph, LangChain, Semantic Kernel, CrewAI, AutoGen or LlamaIndex. • RAG done properly: retrieval strategy, embeddings, vector databases — and diagnosing a weak pipeline. • Model platform fluency: Azure AI / OpenAI APIs, AWS Bedrock or Google Vertex AI. • Cloud-native delivery: Docker, Kubernetes, CI/CD. • Engineering judgement — knowing when a deterministic solution beats a model, and saying so.