How to Apply to Stability AI

9 min read Last updated April 20, 2026 3 current roles tracked
Blake Crosley
Blake Crosley Researched using publicly available information, hiring data, and ATS analysis.

ResumeGeni's employer crawl detects Greenhouse serving Stability AI's application flow across 3 live openings. See how Greenhouse reads your resume.

Key Takeaways

  • Stability AI is a turnaround company, not a hot AI darling — apply with eyes open to the 2024 founder exit, funding crisis, and strategic reset under CEO Prem Akkaraju.
  • Sean Parker as Executive Chairman and Akkaraju's WETA Digital background signal a deliberate pivot toward film, gaming, and creative-industry tools rather than broad consumer AI.
  • The departure of Stable Diffusion's core team to found Black Forest Labs, whose FLUX.1 model has displaced Stable Diffusion in many workflows, is a real competitive overhang.
  • Greenhouse is the ATS — submit clean single-column PDFs, complete all structured fields, and match keywords naturally without padding.
  • Research and engineering bars are high and substance-focused; expect to defend specific model and architecture decisions in detail.
  • Open-source orientation is core identity; informed views on model licensing, copyright, and safety are evaluated explicitly.
  • London is the center of gravity, with smaller San Francisco and remote-UK presence; visa sponsorship is possible but not automatic.
  • Compensation is competitive for the UK market but typically below top-tier US AI labs; equity carries genuine upside-and-downside risk given the turnaround context.

Source basis: This guide combines the company's public careers materials, detected ATS-provider data, and ResumeGeni analysis. Employer-specific details should be read alongside the Sources section below; interview-culture guidance may synthesize public candidate reports when official documentation is limited.

About Stability AI

Stability AI Ltd. is a London-headquartered generative AI company best known for releasing Stable Diffusion in August 2022, the open-source text-to-image model that, alongside DALL-E and Midjourney, kicked off the consumer generative-image era. The company was founded in 2019 by Emad Mostaque, a former hedge-fund analyst and entrepreneur whose public profile became as much a liability as an asset. After the viral release of Stable Diffusion, Stability raised roughly $110 million in a 2022 Series A led by Coatue and Lightspeed Venture Partners at a reported valuation near $1 billion, briefly putting it in the same conversation as OpenAI and Anthropic. That conversation has changed. In March 2024, Mostaque resigned amid mounting questions about the company's finances, governance, and product direction. Christian Laforte and Shan Shan Wong took over as interim co-CEOs, and Sean Parker — co-founder of Napster, founding president of Facebook, and managing partner at Founders Fund — became Executive Chairman. By October 2024, Parker led an approximately $80 million restructuring round, and in June 2024 Prem Akkaraju, the former CEO of WETA Digital (Peter Jackson's visual effects studio behind The Lord of the Rings and Avatar), was appointed CEO. Headcount, reportedly above 200 at peak, has contracted to roughly 150 post-reorganization. Under Akkaraju, Stability is repositioning away from broad consumer AI and toward generative tools for the creative industries: film, television, gaming, and design. His WETA background signals the strategic intent. The product surface is broad: Stable Diffusion 3.5, Stable Diffusion 3 Medium/Large/Turbo, Stable Audio 2.0 for text-to-music, Stable Video Diffusion, Stable LM language models, Stable Cascade, Stable 3D in partnership with Tripo AI, the commercial Stability API, and the consumer DreamStudio app. Most models are released openly, though the 2024 Stability AI Community License Agreement now requires a paid commercial license once an organization passes roughly $1 million in annual revenue. Two overhangs are material to any candidate. First, Getty Images is suing Stability AI in both the UK and the US over the use of its imagery in training data — a case widely watched as a bellwether for generative-AI copyright. Second, in 2024 several core researchers behind Stable Diffusion left to found Black Forest Labs, whose FLUX.1 model has, in many practitioner workflows, displaced Stable Diffusion as the open-weights image model of choice. Stability is a turnaround AI company with serious technology, real revenue ambitions, and significant uncertainty — not a hot AI darling.

Application Process

  1. 1
    Apply through the Greenhouse-hosted careers site (jobs

    Apply through the Greenhouse-hosted careers site (jobs.stability.ai or boards.greenhouse.io/stabilityai). Most listings are London-first with select remote-UK or San Francisco roles.

  2. 2
    Expect a recruiter screen within one to two weeks for active roles

    Expect a recruiter screen within one to two weeks for active roles. Pipelines move slower than at hyper-growth peers because the team is small and selective post-2024 reorganization.

  3. 3
    Technical roles typically include a hiring-manager conversation focused on prior

    Technical roles typically include a hiring-manager conversation focused on prior projects, followed by a technical deep dive — research presentation for scientists, system design or coding for engineers.

  4. 4
    For research and applied research roles, be ready to present a paper, model, or

    For research and applied research roles, be ready to present a paper, model, or open-source project you led, and to defend technical choices in detail.

  5. 5
    Engineering loops generally include a coding round (often Python plus PyTorch or

    Engineering loops generally include a coding round (often Python plus PyTorch or distributed systems), a system-design or ML-infra round, and a values/culture conversation.

  6. 6
    Product, design, and go-to-market roles emphasize portfolio review, especially w

    Product, design, and go-to-market roles emphasize portfolio review, especially work touching creative tools, generative media, or the film, gaming, and design industries Akkaraju is targeting.

  7. 7
    Final-stage candidates typically meet a member of the executive team

    Final-stage candidates typically meet a member of the executive team. Senior hires may meet Prem Akkaraju or Sean Parker depending on the function.

  8. 8
    Decisions are usually communicated within one to two weeks of the final loop

    Decisions are usually communicated within one to two weeks of the final loop. Offers are reference-checked thoroughly given the company's smaller scale and visibility risk.

  9. 9
    Visa sponsorship for the UK is available for senior roles but not guaranteed for

    Visa sponsorship for the UK is available for senior roles but not guaranteed for every position; confirm with the recruiter before investing in the loop.

  10. 10
    Be candid in early conversations about your view of Stability's trajectory, the

    Be candid in early conversations about your view of Stability's trajectory, the Black Forest Labs departures, and the Getty litigation. Recruiters expect informed candidates.

Resume Tips for Stability AI

recommended

Lead with concrete generative AI experience: papers, open-source contributions,

Lead with concrete generative AI experience: papers, open-source contributions, model checkpoints released, or production diffusion or transformer systems you have shipped.

recommended

Quantify model work where possible: parameter counts, training tokens or images,

Quantify model work where possible: parameter counts, training tokens or images, FID or CLIP scores, throughput improvements, or inference cost reductions.

recommended

For research roles, include a publications section with arXiv links and conferen

For research roles, include a publications section with arXiv links and conference venues (NeurIPS, ICML, ICLR, CVPR, SIGGRAPH). HuggingFace model pages and GitHub repositories count as primary evidence.

recommended

Engineering candidates should highlight PyTorch, distributed training (DeepSpeed

Engineering candidates should highlight PyTorch, distributed training (DeepSpeed, FSDP, Megatron), CUDA, Triton, and large-scale data pipelines (Ray, Spark, custom).

recommended

Surface any film, VFX, gaming, animation, or design-tool experience prominently

Surface any film, VFX, gaming, animation, or design-tool experience prominently. Akkaraju's strategy explicitly targets these verticals and that background is a real edge.

recommended

Mention open-source contributions, Discord or community engagement, and any work

Mention open-source contributions, Discord or community engagement, and any work involving model licensing, evaluation, or safety. Stability's open-weights posture is core identity.

recommended

Avoid generic AI buzzwords

Avoid generic AI buzzwords. Specific model names (SDXL, SD3, Stable Audio, FLUX, DiT, latent diffusion) and techniques (LoRA, ControlNet, classifier-free guidance) read as credible.

recommended

Keep the resume to one or two pages

Keep the resume to one or two pages. UK-style CVs are accepted but tighten ruthlessly — recruiters skim, and concise documents perform better in Greenhouse review.

recommended

If you have GPU cluster, training infrastructure, or large-scale inference servi

If you have GPU cluster, training infrastructure, or large-scale inference serving experience (Triton, vLLM, custom kernels), call it out. Compute economics matter at Stability's scale.

recommended

Do not overstate

Do not overstate. The team is small and the AI research community is incestuous; exaggerated claims surface fast in references.

ATS System: Greenhouse

Stability AI uses Greenhouse for applicant tracking. The careers site at jobs.stability.ai and the public board at boards.greenhouse.io/stabilityai both feed the same Greenhouse pipeline. Greenhouse parses standard PDF and Word resumes reliably and scores them against structured job requirements set by hiring managers.

  • Submit a clean, single-column PDF without tables, multi-column layouts, headers, or text boxes — Greenhouse's parser handles linear text best.
  • Match keywords from the job description naturally — model names, frameworks, and tools — but never keyword-stuff. Recruiters review parsed output and notice padding.
  • Fill out the structured Greenhouse application fields completely, including custom questions. Skipped questions are visible to recruiters and read as low effort.
  • Use a single professional email address and consistent name across LinkedIn, GitHub, and HuggingFace so Greenhouse's deduplication and reference tooling works cleanly.
  • If you have prior contact with anyone at Stability, mention it in the referral or 'how did you hear about us' field — internal referrals are weighted in Greenhouse scorecards.
  • Attach a cover letter only when it adds substance: a specific reason you want to work on Stable Diffusion, Stable Audio, or the creative-industries strategy. Generic letters are noise.
  • Apply once per role. Greenhouse flags duplicate applications, and reapplying within ninety days for the same job rarely changes the outcome.

Interview Culture

Interviews at Stability AI reflect a small, researcher-heavy team that has been through a difficult year and is rebuilding under new leadership.

The bar is high on technical substance and low on theatrical performance. Expect direct, fairly informal conversations — UK and ex-academic norms dominate, even for the San Francisco contingent. Interviewers want to see that you have actually built or trained generative models, not just consumed papers about them. Research loops emphasize a paper or project presentation followed by deep technical questions. Be ready to defend architectural choices, explain failure modes, and discuss what you would do differently. Engineering loops are pragmatic: coding rounds in Python with a PyTorch or distributed-systems flavor, system-design rounds focused on training or inference infrastructure at scale, and a debugging or operational round for senior infrastructure roles. Product and design loops center on portfolio walk-throughs with emphasis on creative tooling, professional workflows, and an honest read on where generative AI is and is not ready for production creative pipelines. Values conversations are real. Interviewers probe your view of open-source AI, model safety, copyright and training-data ethics, and the state of the generative-image market post-FLUX.1. Vague or evasive answers do not land well. Stability hires people who have a point of view and can defend it without rigidity. Expect at least one conversation about why you want to join a turnaround AI company specifically — they want candidates with eyes open to the risk, not opportunists chasing a logo. Final rounds typically include time with an executive; at the senior level, this can include Prem Akkaraju or Sean Parker.

What Stability AI Looks For

  • Demonstrated generative AI expertise: shipped models, published papers, or substantial open-source contributions to diffusion, transformer, or multimodal systems.
  • Strong PyTorch and distributed training fundamentals for engineering and research roles, including familiarity with FSDP, DeepSpeed, or Megatron-style parallelism.
  • Conviction about open-source AI and an informed view of the licensing, safety, and copyright tradeoffs Stability navigates.
  • Experience or genuine interest in creative-industry workflows — film, VFX, gaming, music, design — aligned with the Akkaraju-era strategy.
  • Pragmatism about compute economics, inference cost, and the operational realities of serving models at scale on constrained budgets.
  • Comfort with ambiguity and turnaround dynamics. Stability is rebuilding, not coasting; candidates who need a stable, fully resourced environment will struggle.
  • Direct communication and the ability to disagree productively. The team is small enough that political or passive-aggressive operators are quickly visible.
  • For senior roles, evidence of having shipped real products to real users — research-only profiles tend to do better at competitors with larger research budgets.
  • Cultural fit with a London-anchored, remote-friendly team that runs leaner and quieter than a typical Bay Area AI startup.
  • Honest self-assessment of what you know and do not know. Bluffing in interviews is filtered out aggressively.

Frequently Asked Questions

Is Stability AI still a viable company after the 2024 crisis?
Stability survived 2024 through a roughly $80 million restructuring round led by Sean Parker in October 2024, and is operating under new leadership with CEO Prem Akkaraju and a refocused creative-industries strategy. The company is smaller (around 150 employees, down from over 200) and the path forward is genuinely uncertain. It is a real company with real technology and real revenue ambitions, but it is also a turnaround story, not a growth-stage success. Treat the equity and the role accordingly.
Why did Emad Mostaque resign in March 2024?
Mostaque resigned amid sustained reporting on the company's finances, internal governance issues, and disputes over his public statements. Press coverage in 2023 and early 2024 raised questions about cash burn and credibility that ultimately became untenable for investors. Stability's official statement was brief, but the practical effect was a leadership and capital reset led by Coatue and Sean Parker.
How does Stability AI compete with FLUX.1 from Black Forest Labs?
It is a real challenge. Black Forest Labs was founded in 2024 by former Stability researchers central to Stable Diffusion, and FLUX.1 has been adopted widely in open-source workflows where Stable Diffusion previously dominated. Stability is responding by iterating on Stable Diffusion 3.5, expanding into video, audio, and 3D, and repositioning toward creative-industry customers under Akkaraju, where integration, support, and licensing matter as much as raw model quality. Whether that is sufficient is an open question.
What is the Getty Images lawsuit and does it affect hiring?
Getty Images is suing Stability AI in both the UK High Court and the US District Court for the District of Delaware, alleging that Stable Diffusion was trained on copyrighted Getty imagery without license. The cases are ongoing and widely watched as a bellwether for generative-AI copyright. It does not stop hiring but it is a material business and reputational risk that any candidate should understand.
Where is Stability AI headquartered and where will I work?
The company is headquartered in London, with research presence in Oxford and a smaller office in San Francisco. Most roles are London-anchored with remote-UK flexibility for senior positions. US roles exist primarily in San Francisco. The company is not a globally distributed remote-first organization; geography matters.
What ATS does Stability AI use and how should I tailor my application?
Stability uses Greenhouse, accessible at jobs.stability.ai or boards.greenhouse.io/stabilityai. Submit a single-column PDF without tables or text boxes, complete every structured field including custom questions, and match keywords from the job description naturally. Apply once per role; Greenhouse flags duplicate applications and rapid reapplication for the same role rarely changes outcomes.
Does Stability AI sponsor visas?
For UK roles, Stability holds a Skilled Worker sponsor licence and sponsors visas for senior and specialist positions, but not every role qualifies. For US roles in San Francisco, sponsorship varies. Confirm with the recruiter before progressing through the loop, especially for non-senior roles.
What is compensation like at Stability AI?
Compensation is competitive within the UK technology market and reasonable in San Francisco, but typically below top-tier US AI labs such as OpenAI, Anthropic, and Google DeepMind. Equity is granted and carries genuine upside in a successful turnaround scenario, but valuations have reset since the 2022 Series A and outcomes are uncertain. Cash compensation is the more reliable component to evaluate.
What kinds of products and models would I work on?
The active product surface includes Stable Diffusion 3.5 and earlier image models, Stable Audio 2.0 for text-to-music, Stable Video Diffusion, Stable LM language models, Stable Cascade, Stable 3D in partnership with Tripo AI, the commercial Stability API, and the DreamStudio consumer application. Under Akkaraju, expect increasing focus on tools and pipelines aimed at film, gaming, and design professionals.
Is Stability AI still committed to open-source?
Yes, with caveats. Most models continue to be released with open weights, but the 2024 Stability AI Community License Agreement now requires a paid commercial license for organizations above approximately $1 million in annual revenue. Open-source remains core to identity and recruiting appeal, but commercial monetization is now an explicit priority.
What is the interview process timeline?
From application to offer, expect three to six weeks for most roles, sometimes longer for senior positions that involve executive interviews. The pipeline includes a recruiter screen, hiring-manager conversation, two to four technical or portfolio rounds depending on function, and a final values or executive conversation. Decisions are typically communicated within one to two weeks of the final loop.
Should I mention the Black Forest Labs departures in interviews?
If it is relevant to the conversation, yes. Stability's interviewers expect candidates to be informed about the competitive landscape and the company's recent history. Demonstrating that you understand the FLUX.1 dynamic, the leadership reset, and the strategic pivot signals seriousness. Avoiding the topic entirely reads as either uninformed or evasive.

Current Role Context

ResumeGeni currently tracks 3 roles for Stability AI. Use the company profile for current role context before tailoring your resume.

Check Your Resume Before Applying → Review Stability AI role context

Related Resources

Related Articles

Sources

  1. Stability AI Official Site —
  2. Stability AI Careers (Greenhouse) —
  3. Stability AI CEO Emad Mostaque Resigns - Forbes —
  4. Sean Parker Becomes Executive Chairman of Stability AI - Bloomberg —
  5. Prem Akkaraju Appointed CEO of Stability AI - Stability AI Blog —
  6. Stability AI Closes $80M Funding Round - Reuters —
  7. Black Forest Labs Launches FLUX.1 - TechCrunch —
  8. Getty Images v. Stability AI - The Verge —
  9. Stable Diffusion 3.5 Release - Stability AI —
  10. Stability AI Community License Agreement —
  11. Stability AI Coatue Lightspeed Series A - TechCrunch —
  12. Stable Audio 2.0 Launch - Stability AI —