Ai Infrastructure Engineer, Agentic Systems For Biomedical Research
Pubblicato il 13-09-2026 - IRCCS Policlinico San Donato in Bardi
Institute for Molecular and Translational Cardiology (IMTC), IRCCS Policlinico San Donato, San Donato Milanese (Milan, Italy) Fixed-term contract, tempo pieno, on site.
Our group is led by Prof. Luigi Anastasia, Professor of Biochemistry at Universita Vita-Salute San Raffaele.
We work at the intersection of chemical biology, regenerative medicine and cardiovascular research, combining experimental science with omics and computational analysis.
Publications: Scholar:
We are opening a new position to build the software layer that will let AI agents support our research in a dependable way.
This is an engineering role, distinct from the omics and data analysis carried out by our scientists.
The focus is not a specific scientific question, but the infrastructure through which answers are produced: its design, deployment and operation.
Responsibilities
Design and build the infrastructure that lets AI agents reach scientific literature, internal documents, analysis pipelines, databases and computing resources through clearly defined tools and permissions.
Deploy and operate that layer: versioning, logging, monitoring, backup, secrets management and human approval for sensitive actions.
Define how an agent is assessed before it is trusted in routine work: accuracy, reproducibility,
failure modes and cost.
Integrate it with our existing data and analysis environment without weakening security or reproducibility.
Maintain the systems and document them so the team can operate and extend them.
Requirements
Solid Python, demonstrated through academic, personal, open source or professional work.
Practical experience with LLM based applications that goes beyond prompting: tool use, state, error handling, multi step workflows.
Working knowledge of Linux, Git, containers and APIs.
Willingness to own a system after the first prototype, and to test it seriously rather than demonstrate it.
Clear written and spoken English.
Italian is useful but not required.
Desirable, not required: agent frameworks, MCP, self hosted model serving, CI/CD, workflow managers, experience with sensitive or health related data, open source work.
The position suits an early career engineer, including an outstanding recent graduate or a PhD candidate.
We weigh evidence of what you have built more than job titles or years of experience.
Remuneration will be commensurate with experience and qualifications.
Applications are reviewed on a rolling basis.
#J-*****-Ljbffr