Senior Research Engineer

Agent Safety Infrastructure

Bocconi · REALMilan, ItalyFull time · Two years

Call upcoming

Build the Airlock and the infrastructure connecting proof checks to tools, resources, and execution.

Your role

You will shape the systems foundations of verifiable agent autonomy: infrastructure through which flexible agents using learning methods can take consequential actions under explicit commitments. The project aims to make these ideas usable beyond a single demonstration, enabling researchers to investigate different models, tools and interaction patterns on a common execution framework.

This is an opportunity for architectural ownership and technical leadership in an emerging research programme. You will influence how formal guarantees connect to real execution, which abstractions can support a reusable framework, and how others can build on the resulting system. As it matures, there will be scope to help guide the public reference implementation and its developer community.

Within PC2, your main focus will be the Airlock, the execution layer connecting an agent's proposed actions to the evidence required by its contract. The first phase will develop a reusable core library for contractual state, proof requests, execution decisions and recorded outcomes, with clear extension interfaces for new benchmarks, toolsets and agent workflows. You will work with researchers in formal methods and empirical research to identify what should be shared and what belongs in adapters for individual environments.

A second phase will develop stronger execution guarantees, including sandboxing and isolation, controlled access to credentials and resources, attestation and recovery from failures. The initial architecture will define the interfaces and trust boundaries required for this subsequent hardening. You will choose technologies and shape research priorities with the team, establishing the testing, documentation and release practices that allow the infrastructure to be inspected and reused.

Your contribution to the field

Your contribution will be a reusable systems architecture for enforceable agent commitments, connecting formal reasoning to the execution of real tools. Through the Airlock's interfaces, reference implementation and experimental evidence, you will shape how agent developers integrate behaviour governed by proof checks across environments. The framework will provide a basis for independent implementations and further security research, with scope for technical leadership of a public project at the intersection of agent infrastructure and formal verification.

An initial research or development direction

An example starting point is to connect two contrasting tool environments through a common Airlock core and use their differences to test its abstractions, execution traces and failure handling. This establishes a basis for investigating how stronger isolation and attestation change the system's guarantees and costs.

Essential background

Substantial practical software and systems engineering experience, with evidence of architectural ownership of a reusable library, execution system or research platform.
Strong implementation, testing and debugging skills, including the ability to reason about state, interfaces and failure behaviour.
Ability to turn evolving research requirements into maintainable software and to explain architectural decisions to collaborators.
Relevant technical training or equivalent professional experience. A publication record or PhD is not a prerequisite for this engineering profile; formal appointment eligibility will be specified in the final call.

Additional relevant experience

Security engineering, isolation, sandboxing, distributed systems, credential management or attestation.
Agent runtimes, tool integrations, observability or reproducible experimental infrastructure.
Formal methods, dependent type theory, interfaces for proof checking, or experience maintaining an open source project.

Candidates with depth in a relevant area and an interest in connecting it to others are encouraged to apply. Evidence of your contribution can include research papers, formal developments, substantial software or a clear account of technical work that cannot be shared publicly.

Research environment and development

Your primary supervisor will be Alessandro Rudi. PC2 is a joint effort involving PRAXIS at AI4I, the REAL team at Bocconi and STAMP at Inria. The role requires close collaboration with the other participating groups, including opportunities for extended visiting periods at their laboratories, bringing your work into close contact with complementary expertise in machine learning, formal methods and systems.

The position is based in Milan, Italy. Remote working arrangements can be agreed when working from another laboratory in the PC2 collaboration. The timing and practical arrangements for research visits will be discussed with you and the participating hosts.

Computational resources. The joint programme is supported by AI4I's HPC infrastructure: approximately 150 NVIDIA GPUs (80 Blackwell B200 and 68 Hopper H200) and 5 PB of storage for agentic AI training, testing and validation. Access for collaborative projects will be arranged through the participating teams.

Application materials

Please prepare:

Your CV, including relevant research or technical experience.
A short statement of fit and interests, explaining what attracts you to the role, the experience you would bring and directions you would be interested in exploring. A full research proposal is not required.
One or two representative contributions, such as papers, formal developments or software. Briefly explain your own contribution. An account of work that cannot be shared publicly is welcome; a public repository is not required.

Submission instructions and the application deadline will be published here when the call is finalised.