GPU Consultant Engineer

Job Description

• About This Role Build open source graphics drivers that run on real hardware. You will work directly in the Vulkan driver stack, contributing upstream to Mesa and the Linux DRM subsystem, and solving problems that affect how GPUs are used across the Linux ecosystem. This isn't just maintenance work. You will be debugging real hardware behaviour, reading GPU specs and kernel internals and writing code that ends up in distributions used by millions of people. The work is technically deep and the feedback loop is direct. The role combines low-level systems programming with open source collaboration. You will work alongside experienced graphics engineers, develop genuine expertise in how modern GPUs work from the kernel up, and grow into a contributor that the broader open source graphics community knows by name. Location: Fully remote Travel: Optional conference attendance
• What You'll Actually Do Upstream First: Contribute directly to Mesa and the Linux DRM subsystem, writing Vulkan driver code that lands in mainline and ships to users worldwide Real-World Impact: Work on graphics drivers for hardware used across Linux desktops, embedded devices and consumer products that people depend on every day Solve Hard Problems: Debug GPU behaviour that has no documentation, reverse engineer hardware quirks and fix rendering issues that manifest as a corrupted frame with no obvious cause. The GPU did something unexpected and it is your job to find out why Community Collaboration: Work with Mesa and kernel maintainers, participate in code review across the open source graphics community and represent Collabora at conferences like XDC and FOSDEM Client Partnership: Translate hardware vendor and client requirements into upstream-compatible solutions while advocating for approaches that benefit the broader ecosystem. Help clients understand why investing in open source driver quality serves their long-term goals Navigate Constraints: Balance upstream correctness with real-world pressures: hardware bring-up timelines, vendor NDAs, performance targets and client deadlines, while keeping code quality high enough to pass maintainer review Be a Technical Advisor: Explain complex driver and GPU architecture decisions to product teams, provide honest assessments of what is and isn't possible and build trust through transparent communication
• What we’re looking for Required Vulkan - solid understanding of the core spec; command buffers, render passes, synchronisation, memory management and pipelines. Familiarity with Vulkan validation layers and debugging workflow Basic understanding of GPU hardware concepts - queues, tilers vs IMR, memory hierarchies Linux kernel basics - comfortable reading and navigating kernel code DRM fundamentals - GEM/buffer management, dma-buf/prime buffer sharing, fence and sync objects, GPU job submission paths Mesa architecture awareness - Gallium, driver structure Basic understanding of how userspace drivers interact with kernel C and C++ - comfortable with pointer arithmetic, memory management, kernel-style coding
• Nice to have Git - mailing list patch workflow (not just GitHub PRs) Comfortable reading specs and documentation independently Not afraid of large unfamiliar codebases Open source contribution mindset - code review, public communication, upstream first Existing Mesa or kernel patch history Experience with a specific GPU family (ARM/AMD/Intel) Python for tooling/scripting Rust programming skills Understanding of KMS fundamentals or Wayland clients
• What We Offer Compensation & Benefits We offer a competitive benefits package tailored to each country in which we have employees, plus a core global benefit offering accessible to all Collaborans. Our core benefits include: Competitive salary based on experience and location Office setup budget Sabbatical/Retention of Services - after five years continuous service Co-working policy - support for working outside home Wellness assessments - biannual wellbeing assessments with a trained mental health specialist Conference attendance - we cover expenses and encourage speaking opportunities Open source time - contribute to projects you care about Work Environment Fully remote - work where you’re most productive Global team - engineers across Europe, Americas and Asia No on-call rotations - this isn't support work Sustainable pace - we're here for the long term Flexible hours - manage your own schedule Growth & Impact Upstream contributions - your work shapes Linux, not just client products Technical leadership track - path to principal engineer, subsystem expertise or technical architecture roles Consulting track - develop deeper client relationships, lead customer engagements or shape service offerings Mentorship opportunities - both receiving and providing guidance Conference speaking - we support presentation submissions and provide coaching Recognition - your contributions are public, building your professional reputation Culture & Values Upstream first - we believe the best solutions benefit everyone Transparency - open discussion about technical decisions, business constraints and project challenges Pragmatism - perfect is the enemy of good; we ship quality code that solves real problems Respect for expertise - we trust engineers to make technical decisions Collaborative - we succeed by working together, internally and with the community Sustainable pace - we're here for the marathon, not the sprint A Day in the Life Morning (your timezone): Review overnight Mesa GitLab discussion on your patch series Triage a rendering corruption report from a client - identify whether it is a driver bug or a spec interpretation issue Respond to maintainer feedback on a Vulkan extension implementation Pair debug session with a colleague on unexpected GPU hang behaviour Midday: Deep work: implement a missing feature in the Vulkan driver, cross-referencing the spec and hardware documentation Quick video call with client team to walk through your proposed fix and explain the trade-offs Review a colleague's DRM buffer management patch before they send it upstream Help a colleague debug an issue to identify whether it is GPU related Afternoon: Reproduce a client issue locally using RenderDoc and Vulkan validation layers - narrow it down to a specific draw call Update client on timeline, discuss whether a quick workaround or a proper upstream fix is the right move for their schedule Join optional team discussion on how to handle a tricky synchronisation edge case Write a clear commit message explaining not just what changed but why the hardware requires it Some days are heads-down driver work with hours spent in a debugger. Others involve more client communication and architectural discussion. You will find a balance that suits the project and your strengths.
• Our Interview Process We respect your time and provide a clear, structured process: Application Review (within 1 week) - we look at your code and contributions Initial Chat (30 min) - casual conversation about your experience and our work Technical Interview (2 - 3h) Technical Deep-Dive - discuss kernel architecture, past debugging challenges and how you approach problem-solving (no whiteboard coding) Code Review Session - review real patches together, discuss trade-offs Client Scenario Discussion - talk through a realistic customer situation: balancing technical debt, timeline pressure and upstream goals (we want to see your thought process, not a "right answer") Head of Open Source Software Consulting team Discussion (60 min) - meet the Head of Open Source Software Consulting team, ask about day-to-day work and how Collabora works, discuss compensation and benefits. Offer (within 1 week of final interview) Total timeline: 3-4 weeks from application to offer We provide feedback at every stage, and you'll always know where you stand. What Happens Next We'll acknowledge your application within 2 business days Initial response within 1 week Even if we don't move forward, we'll tell you why
• Collabora's Commitment We're committed to building a diverse team and inclusive environment. We encourage applications from people underrepresented in tech. We evaluate candidates based on skills and contributions, not pedigree. If you're unsure whether you qualify, apply anyway —imposter syndrome is real and you might be exactly who we're looking for. We recognise that the combination of deep kernel expertise AND consulting skills is rare. If you're strong in one area and interested in developing the other, we want to hear from you.
• #LI-Remote
• We may use artificial intelligence (AI) tools to support parts of the hiring process, such as reviewing applications, analyzing resumes, or assessing responses. These tools assist our recruitment team but do not replace human judgment. Final hiring decisions are ultimately made by humans. If you would like more information about how your data is processed, please contact us.
• apply for this job