Linux drivers &
userspace integration
Connect kernel interfaces with userspace libraries and command-line tools. Focus on clear APIs, asynchronous behavior, concurrency, and resource lifecycles.
Software Engineer at AMD
Hi, I'm Anushrut.
Building beneath
the surface.
Firmware. Drivers. Networks. I work on the layers you don’t usually see, so the things you do see work reliably. Welcome to my corner of the stack.
SystemsAcross the stack
2,849GitHub contributions
2017Contributing since
GitHub contribution-calendar snapshot · 2017–October 2026 · Public & private activity.
engineer / systems
AnushrutSoftware Engineer · AMD
// from silicon to software
01Infrastructure & automationPython · Ansible
02Userspace & control planeC++ · Go
03Drivers & firmwareC · Linux · RTOS
Less jargon, please
I help hardware and software talk to each other, then build the tools to check they’re getting along. Sometimes that means following one packet a very long way.
01 / About
Curiosity first.
Then, under the hood.
I’m a Software Engineer at AMD, working across networking firmware, Linux device drivers, and userspace systems. I connect platform behavior with the tools engineers use to configure, validate, and troubleshoot it.
My work spans programmable packet processing, embedded RTOS simulation, kernel-to-userspace APIs, and GPU networking infrastructure. I also build diagnostics, monitoring pipelines, and reproducible deployment and test workflows.
In personal projects, I bring that systems mindset to developer tools and AI-assisted workflows. I enjoy making complex work easier to follow, with usable interfaces and a person in the loop.
My kind of rabbit hole? Following a surprising behavior across the stack until it makes sense. The satisfying part is turning that “why?” into a fix, and a test that catches it next time.
Follow the whole path
From a firmware response to a driver interface to an operator’s command line.
Make failures explainable
Useful telemetry, clear error reporting, and diagnostics that lead somewhere.
Validate real behavior
Meaningful assertions, realistic mocks, and hardware-aware integration tests.
How I work
Built with AI.
Thought through by me.
Most of my development is now done with and alongside AI. I use it as a collaborator to explore code, develop implementations, and iterate on tests.
I own the design decisions, review the changes, and validate the behavior. AI is part of the workflow, not a substitute for understanding the system.
02 / Engineering expertise
Different layers.
The same attention to detail.
Experience across the systems stack, from implementation to validation and operations.
Linux drivers &
GPU networking &
platform automation
Build configuration-driven workflows for compute nodes, GPUs, and switches. Automate discovery, driver deployment, configuration, verification, and teardown across hardware and simulated environments.
Firmware &
RTOS simulation
Work across embedded software, host-based simulation, and hardware-aware validation. Make builds reproducible and tests useful, with attention to interface correctness and error handling.
Networking &
packet processing
Implement and validate programmable data-plane behavior: routing lookups, flow offload, traffic policing, and statistics. Build packet-level tests for encrypted traffic, address translation, and network isolation.
Diagnostics &
cluster health
Build health checks and telemetry collection in Python and Go. Analyze RDMA queue state, configuration consistency, connectivity, and software versions, with structured reports and monitoring dashboards.
Test infrastructure &
release engineering
Develop unit, protocol, and end-to-end regression coverage. Improve mock fidelity, negative testing, and CI failure detection; maintain containerized builds, Linux packages, and binary-hardening checks.
Developer tools & AI-assisted workflows
Build tools that make complex work easier to follow. Bring together automation, usable interfaces, and human review, with clear feedback and tests for failure paths. The goal: less friction, without giving up engineering judgment.
03 / Toolkit
What’s in
the toolbox?
Languages & systems
Build, test & operations
Networks & diagnostics
04 / Journey
Built one layer
at a time.
Developer tools, with a human in the loop
Explore AI-assisted workflows and web interfaces, with an emphasis on usability, automation, and testing. Keep human judgment central to how the tools are used.
GPU networking & systems integration
Linux drivers, userspace APIs, control-plane services, and fabric-management tooling. Infrastructure automation and regression coverage, with attention to asynchronous operations, concurrency, and hardware-compatible behavior.
Embedded platforms & observability
RTOS simulation, firmware build and installation workflows, cluster health-checking tools, and monitoring pipelines. RDMA diagnostics, structured output, and automated validation.
Programmable networking & data planes
Packet-processing pipelines and test infrastructure covering routing, flow offload, policing, IPsec, NAT, and statistics. Simulation, build integration, and release maintenance.
Network test automation
Go-based integration tests for overlay networking, segmentation, redundancy, and traffic behavior. Testbed configuration, failure diagnostics, reboot coverage, and CI workflows.
Open source & continued learning
Scoped code and developer-documentation contributions, alongside Python tooling, web applications, image-classification APIs, computer vision, robotics, speech-recognition experiments, and algorithms. Smart India Hackathon 2020 finalist.
05 / Get in touch
Got a good
“what if?”
A tricky systems problem, an idea to explore, or just a hello. You don’t need a fully formed proposal to start a conversation.