Agentic Ai Architect

Job Summary:
You will lead Mindsprint's GenAI and agentic AI architecture, taking agent-based solutions from opportunity to production. You will design systems that combine LLMs, RAG, knowledge graphs, tool use through MCP and enterprise integrations, with governance, security and reliability built in.
Job Description:
Work with business leaders to find high-value AI use cases and shape production-grade solution strategies and adoption roadmaps.
Design agentic AI architectures: single- and multi-agent orchestration, planning, memory, tool calling and human-in-the-loop controls.
Architect RAG and knowledge systems: document ingestion and parsing, chunking, embeddings, vector and hybrid search, and knowledge graphs.
Define how agents connect to enterprise systems and data through MCP servers, APIs and event-driven integration.
Select and evaluate models (OpenAI, Anthropic, open-source via Hugging Face), including fine-tuning where it pays off.
Establish enterprise AI governance: guardrails, evaluation frameworks, observability, cost controls, security and responsible-AI policies.
Build reusable AI frameworks, accelerators and reference architectures that cut delivery effort across engagements.
Lead technical workshops, demos, PoCs and executive presentations; translate AI concepts for technical and non-technical audiences.
Partner with the Data Engineering Architect to make sure data platforms are AI-ready.
Guide and mentor AI engineers, data scientists and solution architects.
Profile Description:
Must-have
15+ years in software, data or solution architecture, with 2+ years designing GenAI or agentic systems that reached production.
Hands-on with LLM application frameworks: LangChain, LangGraph, LlamaIndex, Semantic Kernel or similar.
Strong RAG design experience with vector databases (Azure AI Search, pgvector, Pinecone, Weaviate or similar).
Working knowledge of MCP and tool/function-calling patterns.
Solid Python and API design; cloud AI services on Azure (Azure OpenAI, AI Foundry) or AWS/GCP equivalents.
Experience with LLM evaluation, guardrails and LLMOps.
Strong data architecture background: you understand how data quality and pipelines shape AI outcomes.
Proven executive-facing consulting and pre-sales experience.
Good to have
Knowledge graphs (Neo4j, JanusGraph) and semantic search (Solr, Elasticsearch).
Document AI tooling (Apache Tika, GROBID, Azure Document Intelligence).
Fine-tuning and serving open-source models.
Multi-agent frameworks (AutoGen, CrewAI, Agent SDKs).
Domain experience in supply chain, agri/food, retail, insurance or healthcare.
Published accelerators, patents, talks or innovation awards.

Work arrangement
Hybrid