21 ott - Pavia
Msg Global
Overview In this role you will architect and operationalize production-grade AI systems that transform business processes. You’ll translate domain knowledge from business experts into reliable agents, pipelines, and evaluation frameworks. You work closely with cross-functional teams to ensure accuracy, efficiency, and governance of AI solutions in real work. You enable scalable, end-to-end AI capabilities that businesses can trust and adopt, driving measurable impact. Retribuzione / Benefits
* competitive salary
* broad range of benefits Responsabilità
* Decompose business processes into agent-driven workflows and define steps, tools, handoffs, and human-in-the-loop decision points
* Design and build retrieval pipelines (chunking, indexing, embedding selection, hybrid search, reranking, grounding)
* Treat prompt/context engineering as an engineering discipline with versioning and tests
* Build and maintain evaluation layers (golden datasets, automated scoring, model-graded evals, regression tests)
* Instrument production systems for observability (traces, token cost, latency, failure rates, data/model drift)
* Own end-to-end quality and diagnose failures by tracing issues to retrieval, reasoning, tool design, or data
* Choose and manage model trade-offs between quality, cost, and latency as models evolve
* Collaborate with business experts to extract domain logic and turn it into testable artifacts
* Take agents to production with infrastructure teams and ensure responsible live behavior Requisiti fondamentali
* 5+ years of software development experience
* 2+ years with production-ready LLM-based systems
* Strong Python skills and standard engineering practices
* Experience with async, FastAPI, Pydantic
* Hands-on with agent orchestration frameworks (LangGraph, LlamaIndex Workflows, OpenAI Agents SDK, Temporal)
* Experience serving models via managed platforms (AWS Bedrock, Azure AI Foundry, Google Vertex AI) or provider APIs
* Production vector search experience (pgvector, Qdrant, Pinecone, Weaviate, Milvus, Azure AI Search)
* Deep retrieval knowledge (chunking, metadata filtering, query rewriting, hybrid search, reranking)
* Fluency with tool/function calling, structured output, multi-step agent design
* Hands-on with MCP (Model Context Protocol): servers, tool interfaces, authentication, permissions, error handling
* Experience with evaluation tooling (Langfuse, LangSmith, Phoenix; regression suites)
* Ability to work in ambiguity and extract requirements with business experts
* Fluent English
* collaboration with cross-functional teams
* strong problem-solving and analytical thinking
* clear communication of technical concepts to non-technical stakeholders
* Python
* FastAPI
* Pydantic
21 ott - Viterbo
Humangest S. P. A.
21 ott - Urbino
Humangest S. P. A.
21 ott - Milano
Staff
21 ott - Gonnesa
Società riconosciuta