Systems & security research
Adrien Ghosn.
Building secure systems
from the right primitives.
I'm a Senior Researcher at Microsoft Azure Research in Cambridge, working on virtualization, confidential computing, and how to manage trust in cloud deployments.
I'm particularly interested in composable isolation and system design: finding simple primitives that compose to provide strong security guarantees. A recurring goal is to decouple confidentiality from trust in the functional correctness of workloads that process sensitive data—AI agents are a motivating example. In Tyche, we explore this through a Rust-based security monitor that lets mutually distrustful components create, compose, and nest isolation boundaries on commodity hardware, from individual libraries to virtual machines. MICA, accepted at ATC 2026, takes a complementary approach to confidential cloud workflows by controlling and attesting how confidential components communicate.
I completed my PhD at EPFL with Edouard Bugnion, James R. Larus, and Mathias Payer. My doctoral research explored isolation through programming abstractions, extending Go and Python to restrict untrusted libraries and protect confidential computations. Along the way (and before LLMs), I've written operating systems from scratch, Linux kernel drivers, hypervisors, virtualization stacks, compilers, and ELF parsers, linkers, and loaders.
01 / Research
Publications
2026
2024
2023
2021
2019
Publication metadata from DBLP. Preprints and software artifacts are labeled separately; edited proceedings are listed under community service.
02 / Background
Curriculum vitae
Research & industry experience
Microsoft Research, Cambridge / Senior Researcher
November 2021-present / Cambridge, UK
Virtualization, confidential computing, and trust relationships in cloud deployments. I joined Microsoft Cambridge as a postdoctoral researcher, working on isolation for commodity hardware and the Verona infrastructure programming language.
Google, Asylo team / Research internship
June-August 2019 / Kirkland, USA
Explored support for higher-level programming languages in Intel SGX enclaves and built a prototype for executing such code inside SGX. Supervised by Matt Gingell.
ABB Corporate Research / Master's internship
February-August 2015 / Baden, Switzerland
Aperiodic-event support in FASA, including fixed-priority servers, data-driven events, and real-time control applications. Work spanning kernel design, dynamic linking and loading, and software updates. Supervised by Manuel Oriol.
Education
EPFL / PhD in Computer Science, 2016-2021
Lausanne, Switzerland
Trust as a Programming Primitive. Research in the Datacenter Systems Laboratory on programming abstractions for compartmentalization and confidential computing, with Edouard Bugnion, James R. Larus, and Mathias Payer.
EPFL / Master's degree, 2013-2016; Bachelor's degree, 2010-2013
Foundations of Software specialization during the Master's degree.
Northeastern University / Master's thesis, September 2015-March 2016
Boston, USA
Efficient runtime deoptimization for R with Jan Vitek in the Programming Languages Laboratory, covering speculative optimization, on-stack replacement, and LLVM-based JIT compilation.
Carnegie Mellon University / Undergraduate exchange, August 2012-July 2013
Pittsburgh, USA
Dean's List, School of Computer Science. Built an x86 Unix-like kernel and a thread library as part of Operating Systems Design and Implementation.
Selected research & systems work
- Tyche / A Rust isolation monitor providing composable compartmentalization and confidential computing on commodity hardware.
- Enclosure / Extensions to Go and Python for restricting untrusted libraries, with isolation enforced through Intel VT-x and Intel MPK.
- Secured Routines / Extensions to Go for executing goroutines inside Intel SGX enclaves.
- Light-Weight Contexts in Dune / Intra-address-space isolation, memory snapshots, and process virtualization.
- Scalameta & Obey / Typed-AST persistence and compiler-enforced, user-defined linting rules for Scala.
Technical skills
Languages / Rust, Go, C/C++, Python, Java, assembly, and shell scripting.
Systems / Operating systems, virtualization, KVM, Intel VT-x, Intel MPK, trusted execution environments, compilers, language runtimes, and virtual machines.
Supervision & teaching
Semester-project supervision on a Go Intel MPK library and system-call interposition in Go and Python runtimes.
Teaching assistant at EPFL, 2015-2020: Concurrent Programming, Introduction to C and Java Programming, Systems for Data Science, Introduction to Operating Systems, and Functional Programming.
The downloadable PDF is the existing June 2023 CV; the online summary includes the updated title from my institutional profile.
03 / Community
Community service
Conference & workshop organization
SysTEX 2026 / Program committee co-chair
9th Workshop on System Software for Trusted Execution, co-located with EuroSys.
Organizers & committee
Reviewing & program committees
| Year | Conference or workshop | Role |
|---|---|---|
| 2026 | EuroSys | Program committee |
| 2026 | SOSP posters | Poster session program committee |
| 2023 | SoCC | Program committee |
| 2022 | ASPLOS | Program committee |
| 2022 | SysTEX | Technical program committee |
| 2019 | USENIX ATC | External reviewer |
Conference names link to the published committee lists. ASPLOS and SoCC use the conference's public submission-site PC pages; ATC uses the official proceedings' reviewer list.
Other contributions
ECOOP 2016 / Student volunteer
Teaching at EPFL / Teaching assistant, 2015-2020
Concurrent Programming, Introduction to C and Java Programming, Systems for Data Science, Introduction to Operating Systems, and Functional Programming.