How to Apply to Graphcore

Key Takeaways

Tailor every application to the specific Graphcore role by echoing the exact technical terminology from the job description — Greenhouse keyword matching directly influences whether a recruiter finds your profile
Quantify your engineering impact using metrics that Graphcore engineers care about: PPA improvements, verification coverage percentages, performance throughput gains, silicon bringup timelines, or model accuracy improvements
Research Graphcore's IPU architecture before interviewing — read their published papers, explore the Poplar SDK documentation, and understand how their Bulk Synchronous Parallel model differs from GPU execution models
Prepare to demonstrate depth over breadth in technical interviews: Graphcore values engineers who can go five levels deep into a problem, not those who can recite surface-level knowledge of twenty different topics
Highlight any cross-functional collaboration experience on your resume and in interviews, as building a full-stack AI processor requires tight coordination between silicon, firmware, software, and research teams
If you are applying from outside the UK, clearly state your work authorization status and willingness to relocate early in the application — Graphcore operates across multiple countries and needs to assess visa logistics upfront
Apply to roles that genuinely match your seniority level — Graphcore's Staff/Senior Staff/Principal hierarchy carries specific expectations around technical leadership, architectural ownership, and team influence

About Graphcore

Graphcore is a pioneering semiconductor company that designs Intelligence Processing Units (IPUs) — a new category of processor built from the ground up for artificial intelligence and machine learning workloads. Founded in 2016 by Nigel Toon and Simon Knowles in Bristol, UK, Graphcore has positioned itself as a bold challenger in the AI accelerator market, competing against established giants like NVIDIA. In 2024, SoftBank Group acquired Graphcore, providing substantial backing and signaling renewed ambitions in next-generation AI hardware. The company's IPU architecture takes a fundamentally different approach to AI computation, featuring massively parallel processing, a unique Bulk Synchronous Parallel execution model, and in-processor memory designed to handle the irregular, sparse workloads common in cutting-edge AI models. Graphcore's engineering culture is deeply technical and research-driven, attracting world-class talent in silicon design, compilers, systems software, and AI research. With offices spanning Bristol, London, Cambridge, Bengaluru, Oslo, and other locations, the company offers a global yet close-knit working environment where engineers work at the frontier of processor architecture. People want to work at Graphcore because it offers a rare opportunity to shape the future of AI hardware rather than incrementally improving existing designs. The approximately 175+ open roles reflect heavy investment across the full silicon lifecycle — from logical and physical design to embedded firmware bringup, verification, manufacturing test, and applications-layer AI research. If you thrive on solving foundational engineering challenges and want your work to directly influence how AI is computed worldwide, Graphcore represents one of the most compelling opportunities in the industry.

Application Process

1
Explore Roles on Graphcore's Greenhouse-Powered Careers Page
Graphcore hosts all open positions on its careers page powered by Greenhouse. Browse roles by location (Bristol, Bengaluru, Cambridge, etc.) and function (Silicon Design, Software, Research, Manufacturing). Pay close attention to the seniority keywords — Graphcore uses a clear hierarchy including Staff, Senior Staff, Principal, and Lead designations that indicate scope and impact expectations.
2
Submit Your Application Through Greenhouse
Complete the Greenhouse application form, uploading your resume and filling in any required fields such as work authorization, location preferences, and relevant experience. Graphcore's roles are highly specialized, so the application typically includes role-specific screening questions about your experience with relevant technologies — for example, RTL design tools for silicon roles or embedded RTOS experience for firmware positions. Take these questions seriously, as they commonly serve as initial screening filters.
3
Recruiter Screening Call
A Graphcore talent acquisition team member will typically conduct an initial phone or video screen lasting 20-30 minutes. Expect questions about your motivation for joining a specialist AI hardware company, your relevant domain expertise, and logistical factors like location and notice period. This is also your chance to ask clarifying questions about team structure and the specific project you would join.
4
Technical Assessment or Take-Home Challenge
Many Graphcore engineering roles include a technical assessment tailored to the discipline. Silicon design candidates may face RTL coding or verification challenges, embedded firmware candidates may work through low-level debugging or bringup scenarios, and software roles may involve systems programming tasks. Research roles often require presenting a published paper or completing a research-oriented problem set. These assessments are designed to evaluate depth, not breadth.
5
Technical Interview Panel
Graphcore commonly conducts one or more in-depth technical interviews with engineers from the hiring team. For silicon roles, expect discussions around microarchitecture trade-offs, timing closure, or verification methodology. For software roles, anticipate deep dives into compiler design, memory management, or distributed systems concepts. Interviewers typically care as much about your reasoning process and how you handle ambiguity as they do about arriving at a correct answer.
6
Hiring Manager and Cross-Functional Interviews
You will likely meet the hiring manager for a conversation focused on your career trajectory, leadership style (for senior roles), and alignment with the team's current priorities. At Graphcore, where cross-functional collaboration between silicon, firmware, and software teams is essential, you may also speak with engineers from adjacent teams to assess how effectively you communicate across discipline boundaries.
7
Offer and Onboarding
Following successful interviews, Graphcore's recruiting team will extend a formal offer through Greenhouse. Given the specialized nature of Graphcore's work, onboarding commonly includes ramping up on the IPU architecture, the Poplar software stack, and internal design tools. Candidates relocating internationally — particularly to Bristol — can expect the company to provide guidance on visa and relocation logistics.

Resume Tips for Graphcore

critical

Lead with AI Hardware or Processor-Relevant Experience

Graphcore builds processors from scratch, so your resume should immediately signal relevant domain experience. If you have worked on ASIC or FPGA design, custom silicon, AI accelerators, GPU architectures, or high-performance computing hardware, place this prominently in your summary and experience sections. Even if your experience is with a competitor's architecture, Graphcore values the transferable understanding of datapath design, memory hierarchies, and parallel computing paradigms.

critical

Use Graphcore's Technical Vocabulary and Industry Terminology

Greenhouse parses resumes for keyword relevance, and Graphcore's recruiters search for specific technical terms. Mirror the language from the job description precisely — terms like 'RTL,' 'SystemVerilog,' 'UVM,' 'silicon bringup,' 'board-level debug,' 'compiler optimization,' 'MLIR,' 'graph-based computation,' and 'embedded RTOS' carry significant weight depending on the role. Avoid generic phrasing like 'hardware development' when you can specify 'physical design implementation targeting advanced process nodes.'

critical

Quantify Impact on Chip-Level or System-Level Metrics

Graphcore engineers obsess over performance, power, and area (PPA) trade-offs. Wherever possible, quantify your achievements in these terms: 'Reduced critical path timing by 15% through pipeline restructuring,' 'Achieved 98% functional coverage across verification plan,' or 'Optimized kernel execution to deliver 2.3x throughput improvement on sparse models.' These metrics speak Graphcore's language and demonstrate the engineering rigor they seek.

recommended

Highlight Cross-Disciplinary Collaboration

Building a novel processor architecture requires deep collaboration between silicon designers, firmware engineers, compiler developers, and AI researchers. If you have experience bridging these domains — for example, working with a silicon team on hardware-software co-design or collaborating with ML researchers to optimize operator implementations — make this collaboration explicit. Graphcore values engineers who can operate effectively across the full stack.

recommended

Showcase Experience with Full Product Lifecycle or Tape-Outs

Graphcore has brought multiple generations of IPU to market, and they value candidates who have shipped real silicon or production-grade systems. Highlight any involvement in tape-outs, silicon bringup campaigns, post-silicon validation, or production ramp. Mention specific process nodes, EDA tools (Synopsys, Cadence, Mentor), or manufacturing test methodologies you have used. Even for software roles, experience deploying systems at production scale is highly relevant.

recommended

Include Publications, Patents, and Open-Source Contributions

For research scientist and principal-level roles, Graphcore places significant weight on your intellectual contributions to the field. List relevant publications from conferences like NeurIPS, ICML, ISCA, MICRO, or DAC. Include granted patents or pending applications related to processor architecture, AI systems, or compiler technology. Open-source contributions to projects like LLVM, MLIR, PyTorch, or TensorFlow also demonstrate the kind of community engagement Graphcore respects.

recommended

Keep Formatting Clean for Greenhouse Parsing

Greenhouse's resume parser handles standard formats well but can struggle with multi-column layouts, embedded tables, headers/footers containing critical information, and heavily designed templates. Use a single-column format with clearly labeled section headers (Experience, Education, Skills, Publications). Submit as PDF to preserve formatting, and ensure your name, email, and phone number appear in the main body text — not in a header that the parser might skip.

nice_to_have

Demonstrate Awareness of IPU Architecture or Competitive Landscape

If you have direct experience with Graphcore's IPU, Poplar SDK, or PopART framework, mention it explicitly — this is rare and highly valued. Even if you haven't used Graphcore technology directly, demonstrating knowledge of the AI accelerator competitive landscape (NVIDIA GPUs, Google TPUs, Intel Gaudi, AMD Instinct, custom silicon from hyperscalers) shows you understand the market Graphcore operates in and why their architectural approach is differentiated.

ATS System: Greenhouse

Graphcore uses Greenhouse as its applicant tracking system to manage the full hiring pipeline from application intake through offer. Greenhouse enables structured hiring with customizable scorecards, ensuring each candidate is evaluated against consistent, role-specific criteria. Applications are parsed, organized by stage, and visible to hiring teams through a collaborative review interface.

Submit your resume as a PDF with a single-column layout to ensure Greenhouse's parser accurately extracts your experience, skills, and education without formatting errors
Mirror exact keywords and phrases from the Graphcore job description — Greenhouse allows recruiters to filter and search candidates by specific technical terms like 'SystemVerilog,' 'silicon bringup,' or 'ML compiler'
Complete every field in the application form thoroughly, including optional questions — incomplete applications may be deprioritized during recruiter triage in Greenhouse
Avoid using images, graphics, charts, or icons in your resume as Greenhouse's parser cannot extract text from visual elements
Place your most critical technical skills and role-relevant experience in the top third of your resume, as this is the portion most likely to be reviewed first when recruiters preview applications in Greenhouse
If reapplying to a different Graphcore role, submit a freshly tailored resume — Greenhouse retains your previous applications and hiring teams can see your full history with the company
Use standard section headings (Professional Experience, Education, Technical Skills) rather than creative alternatives, as this improves parser accuracy and ensures your information maps to the correct fields

Complete Greenhouse Resume Guide →

Interview Culture

Graphcore's interview process reflects its identity as a deep-tech semiconductor company where engineering excellence is non-negotiable.

Expect a rigorous, multi-stage process that is technically demanding but respectful of your time and expertise. The overall process typically spans three to five rounds, depending on seniority, and may take two to four weeks from initial screen to offer. For silicon design roles (logical design, physical design, verification), interviews commonly involve whiteboard-style or screen-shared exercises exploring RTL design trade-offs, timing analysis scenarios, or verification planning. Interviewers are typically senior engineers or leads who are deeply embedded in IPU development, and they will probe your understanding of real-world constraints — expect questions about how you handle timing violations at advanced nodes, debug silicon failures during bringup, or optimize verification closure. Embedded firmware and software roles focus on systems-level thinking: memory models, concurrency, low-level optimization, and hardware-software interaction. For Principal and Architect-level positions, expect architectural discussion sessions where you may be asked to propose or critique a system design, defend trade-offs, and demonstrate judgment about complexity management. Research scientist interviews at Graphcore typically include a presentation of your recent work (often a published paper), followed by deep technical Q&A from a panel of researchers. They evaluate not just the quality of your research but your ability to connect theoretical work to practical implementation on real hardware. Culturally, Graphcore interviewers value intellectual honesty over performative confidence. Saying 'I don't know, but here's how I would approach it' is far more respected than bluffing. The company's relatively flat engineering structure means you may interview with people who will be your direct collaborators, not distant managers. Demonstrating curiosity about the IPU architecture, asking insightful questions about Graphcore's technical roadmap, and showing genuine enthusiasm for pushing the boundaries of AI compute will resonate strongly. Interviewers often note that candidates who engage in two-way technical discussion — rather than simply answering questions — stand out.

What Graphcore Looks For

Deep domain expertise in at least one core discipline — silicon design, verification, embedded firmware, compiler engineering, or AI/ML research — with enough breadth to collaborate across team boundaries
Hands-on experience with full-product-lifecycle chip development, including tape-out, bringup, validation, and production, not just pre-silicon simulation
Strong first-principles thinking and the ability to reason about complex system-level trade-offs involving performance, power, area, and programmability
Comfort with ambiguity and novel problems — Graphcore builds a fundamentally new processor architecture, so many challenges lack established playbooks
Genuine passion for AI hardware and an understanding of why purpose-built accelerators matter in the evolving AI compute landscape
Effective communication skills, particularly the ability to explain complex technical concepts to colleagues in adjacent disciplines (e.g., a silicon designer explaining constraints to a compiler engineer)
Evidence of ownership and initiative — shipping products, driving technical decisions, mentoring junior engineers, or contributing to the broader technical community through publications or open source
Alignment with a startup-like engineering culture within a now SoftBank-backed organization, meaning high autonomy paired with ambitious delivery expectations

Frequently Asked Questions

How long does the Graphcore hiring process typically take from application to offer?

Based on common patterns at semiconductor companies of Graphcore's size and complexity, the process typically takes two to five weeks from initial application to offer, though this can vary based on role seniority and the team's urgency. The recruiter screen usually happens within one to two weeks of applying, followed by technical interviews scheduled over the following two to three weeks. Principal and Lead-level roles may involve additional rounds, including architectural discussions and cross-functional interviews, which can extend the timeline. Staying responsive to scheduling requests and being flexible with interview times — especially across time zones if you are applying to a different office — can help accelerate the process.

Does Graphcore require a cover letter with applications?

Graphcore's Greenhouse application forms do not always require a cover letter, but submitting one can meaningfully differentiate your application for competitive roles. If you choose to write one, focus on why you are specifically drawn to Graphcore's IPU approach — not generic interest in AI or semiconductors. Reference specific technical challenges the company faces that align with your expertise, such as scaling next-generation silicon, optimizing sparse computation, or building a production-quality software stack for novel hardware. Keep it concise (250-350 words) and use it to provide context that your resume alone cannot convey, such as explaining a career transition or highlighting a particularly relevant project.

What resume format works best with Graphcore's Greenhouse ATS?

Submit a single-column PDF resume with standard section headings and no graphics, tables, or multi-column layouts. Greenhouse's parser handles clean, text-based PDFs reliably, but can misinterpret complex visual designs, sidebars, or icon-based skill ratings. Ensure your contact information appears in the body of the document rather than in headers or footers. For technical roles at Graphcore, a two-page resume is perfectly acceptable and often expected — senior engineers with deep experience in silicon design, verification, or systems software should not artificially compress their qualifications to fit one page.

What technical skills should I emphasize when applying to Graphcore silicon roles?

For logical design roles, emphasize proficiency in SystemVerilog or Verilog, experience with synthesis and timing analysis, and familiarity with microarchitectural design for high-performance processors. For physical design, highlight expertise in place-and-route tools (ICC2, Innovus), static timing analysis (PrimeTime), and experience at advanced process nodes (7nm and below). Verification roles should showcase UVM methodology, constrained random verification, formal verification tools, and coverage-driven approaches. Across all silicon roles, experience with scripting languages (Python, Tcl, Perl) for automation and any exposure to processor or accelerator design — rather than purely FPGA or SoC integration work — will position you strongly for Graphcore's IPU-focused teams.

Does Graphcore hire remote employees, or is on-site presence expected?

Graphcore's approach to remote work varies by role and location. Many semiconductor engineering roles — particularly those involving silicon bringup, board-level debug, and lab-based validation — typically require on-site presence due to the hands-on nature of the work. Software, research, and design roles may offer more flexibility, and Graphcore has historically supported distributed teams across Bristol, Cambridge, London, Bengaluru, Oslo, and other offices. Check each specific job listing on Greenhouse for location requirements, and be prepared to discuss your flexibility during the recruiter screen. Given SoftBank's acquisition, the company's remote work policies may continue to evolve.

Can I apply to multiple Graphcore roles simultaneously?

Yes, you can submit applications to multiple Graphcore roles through Greenhouse, and this is reasonable if you genuinely qualify for more than one position. However, be strategic — applying to ten roles across unrelated disciplines signals unfocused interest and may weaken your candidacy. Greenhouse retains your full application history, so recruiters and hiring managers can see every role you have applied for. If you are genuinely qualified for two or three related positions (for example, both a Senior Embedded FW Engineer and a Principal Embedded FW Engineer), tailor each resume and application to the specific role's requirements and seniority expectations.

What level of experience does Graphcore expect for Senior Staff and Principal-level roles?

Graphcore's Senior Staff and Principal designations typically correspond to seasoned engineers with 10-20+ years of deep domain experience and demonstrated technical leadership. Senior Staff engineers are commonly expected to own significant technical subsystems, make architectural decisions with team-wide impact, and mentor more junior engineers. Principal-level roles carry even greater scope — think company-wide architectural influence, external technical representation, and the ability to define technical strategy for an entire product area. These are not purely managerial roles; Graphcore expects hands-on technical contribution combined with leadership. If you are earlier in your career, targeting Staff or Senior-level roles may be more appropriate.

How should I prepare for a Graphcore research scientist interview?

Graphcore's research scientist interviews typically include a presentation of your recent work, so select a paper or project that demonstrates both technical depth and practical relevance to AI hardware or model development. Be prepared to explain not just what you did, but why it matters for real-world AI systems and how it could translate to implementation on specialized hardware like the IPU. The panel will likely include senior researchers who will probe assumptions, methodology, and scalability. Familiarize yourself with Graphcore's published research on topics like AI model training on IPUs, sparse computation, and their Poplar graph framework. Demonstrating awareness of the computational bottlenecks in modern AI workloads — and how novel architectures might address them — will set you apart.

What is the best way to follow up after submitting a Graphcore application?

After submitting through Greenhouse, you will receive an automated confirmation email. If you have not heard back within two weeks, a polite follow-up email to Graphcore's recruiting team is appropriate — reference the specific role title, requisition number, and briefly restate your strongest qualification. Connecting with Graphcore engineers or recruiters on LinkedIn can also increase your visibility, especially if you engage thoughtfully with content they share about IPU technology or company milestones. Avoid sending multiple follow-ups in quick succession, as this can have the opposite of the intended effect. If you have a referral from a current Graphcore employee, ensure they submit the referral through Greenhouse's internal referral system, as this typically flags your application for prioritized review.

Sample Open Positions

Check Your Resume Before Applying → View 175 open positions at Graphcore

Related Resources

Similar Companies

Sources

Graphcore Careers Page — Graphcore
Graphcore Company Overview and IPU Technology — Graphcore
SoftBank Completes Acquisition of Graphcore — Graphcore
Graphcore Reviews and Interview Experiences — Glassdoor
Greenhouse ATS Candidate Help and Resume Formatting — Greenhouse Software