01ENGINEERING
Build solid systems
Care about the implementation: what runs, where it breaks, and what it costs.
implementation / performance / failure cases
BUILD / EXPLORE / SHARE
Yanwen
Xia.
夏彦文 Beijing, China
Systems, tools, and things worth sharing.
I like building systems that people can run, understand, and build on. I aim to share the code, experiments, and lessons whenever possible.
Amazed by the power of AI agents, I’m exploring what I can build with them.
My systems background is in far memory, fault tolerance, and runtimes.
Currently a PhD student at ICT, CAS & UCAS.
$ whoami
Yanwen Xia / 夏彦文Systems, tools & curious exploration
$ cat approach.txt
build solid systems · share what I can
checkable experiments · shared credit
$ ls work/
profile.sessionUTF-8 / en + zh
A small terminal. Try help, work, or fortune.
BUILD / INSPECT / SHARE / COLLABORATE
01 / APPROACH
Built to be useful.
How I want to work,
and what I bring to a project.
I can help with systems prototyping, performance debugging, and setting up experiments that others can inspect.
02OPEN SOURCE
Share usable work
When I can share the work, I aim to include source code, setup steps, and evaluation scripts.
code / documentation / runnable examples
03REPRODUCIBILITY
Make results checkable
Be clear about the environment, assumptions, and limits, so others can inspect and repeat the experiments.
environments / experiments / limitations
04COLLABORATION
Build with people
Be clear about my part, credit collaborators, and leave code and notes that someone else can pick up.
shared ownership / clear contributions / useful handoffs
02 / WORK
Things I've worked on.
These systems are research collaborations. Each entry describes my part and links to the full team and available materials.
Research collaboration
Starfish @ATC 2026
A fault-tolerant far-memory runtime designed to keep protection overhead low.
My partBuilt the C++ research prototype from scratch and led the project end to end.
Public code, setup instructions, and experiment scripts. Running the system requires a multi-server environment with high-speed networking.
Research collaboration
Orthrus @SOSP 2025
Detecting silent data corruption in cloud applications through computation validation.
My partBuilt the LLVM-based fault-injection infrastructure and an automated Python fault-injection platform.
Public code, environment setup, and evaluation instructions are available in the linked repositories.
Research collaboration
Beehive @NSDI 2025
A far-memory runtime that overlaps remote-memory access with asynchronous execution.
My partImplemented object-based runtime memory management. Evaluated Llama, NAS MG, and SocialNet workloads, and contributed the motivation analysis and figures.
The repository includes C++ and Rust implementations, build/run instructions, and benchmark data references.
Research collaboration
MoleculeGC @EuroSys 2027
Garbage collection for tiered memory, with short pauses and efficient CPU use.
My partLed the experimental evaluation and implemented comparison baselines. Contributed to the design and evaluation write-up.
Project summary and full team below.
Other hands-on work
I've also worked on an xv6-based RISC-V kernel and scientific-computing optimization, and led a student supercomputing team.
Engineering background ↗03 / AWAY FROM THE KEYBOARD
Still curious.
I like exploring systems beyond the screen, too.
A recent favorite TV series is 大明王朝1566.
04 / CONTACT
Let's build something useful.
Happy to compare ideas, work through systems problems, or explore agent tools together.
yenwenshare@gmail.com
Other ways to reach me
Qzone: 2437103317
WeChat: 18252677381