About Thinking Machines

The mission of Thinking Machines is to build AI that extends human will and judgment. We are training frontier models with Inkling, developing Tinker to let people make models their own, and crafting interfaces that broaden human-AI communication. We believe the future worth building is human, and we're hiring people who want to build it.

About Tinker & Inkling

Tinker is our fine-tuning API that empowers researchers and developers to customize frontier AI to their needs — opening access to capabilities that have previously been concentrated in a handful of labs. We manage the infrastructure while allowing Tinkerers full flexibility in training open weights models with their own data, algorithms, and for their own needs.

Inkling is our open-weights model. It is a Mixture-of-Experts transformer with 975B total parameters, 41B active. It supports a context window of up to 1M tokens. It was pretrained on 45 trillion tokens of text, images, audio and video. It is the first in a family of models of different sizes.

About the Role

We're hiring a Developer Relations Engineer to help the open source community build with Thinking Machines' products — Inkling and the models that follow it, and Tinker, our fine-tuning platform. You'll work hands-on with real users, turn their feedback into product improvements, and help developers get the most out of our models and tools.

This role sits at the intersection of engineering, research and community: you'll write code, ship industry leading ML recipes and integrations, and also represent Thinking Machines externally through talks, demos, and technical writing. It's a great fit for someone who's happiest building in a code editor, but who also enjoys teaching and engaging with a fast-growing developer community.

What You'll Do

Write, test, and maintain high-quality agentic training, code examples and documentation for Inkling and Tinker
Work directly with developers and researchers in the open source community to debug issues, gather feedback, and improve their experience fine-tuning and deploying models
Partner with the research and infrastructure teams to turn community patterns and feedback into concrete improvements to Inkling, future models, Tinker, and develop new post-training recipes.
Build demos, integrations, and tools that make it easier for developers to adopt and get value from Thinking Machines' models
Represent Thinking Machines in the open source community through blog posts, talks and tutorials, livestreams, and conference and meetup appearances
Help shape the developer experience roadmap as new models and platform features ship

Skills & Qualifications

Minimum Qualifications

5+ years of experience in developer relations, developer experience, or a software engineering role with significant open source or community-facing work
Strong software engineering skills, with experience writing and shipping production-quality Python code
Hands-on experience fine-tuning, evaluating, or deploying large language models

Preferred Qualifications

Experience with supervised and reinforcement learning fine-tuning of LLMs, post-training on new domains, and improving model capability
Track record contributing to, maintaining, or growing open source software libraries
Experience writing public or internal tutorials, cookbooks, or technical content that helped others understand and apply technical concepts
Comfortable engaging publicly through talks, livestreams, or conference and meetup appearances
Familiarity with multimodal models (text, image, audio) and the tradeoffs involved in training, fine-tuning, and inference

Logistics

Location: This role is based in San Francisco, California.
Compensation: Depending on background, skills and experience, the expected annual salary range for this position is $300,000 - $350,000 USD.
Visa sponsorship: We sponsor visas. While we can't guarantee success for every candidate or role, if you're the right fit, we're committed to working through the visa process together.
Benefits: Thinking Machines offers generous health, dental, and vision benefits, unlimited PTO, paid parental leave, and relocation support as needed.

About the company

Thinking Machines Lab researches and builds AI systems, aiming to make AI more widely understood, customizable, and capable. It publishes much of its research openly on its Connectionism blog, starting with work on making AI models give consistent answers to the same question. Its first product, Tinker, lets researchers and developers fine-tune open-weight AI models on their own data while Thinking Machines runs the computing clusters behind it, so customers focus on their data and training methods instead of managing hardware and pay for the computing they use. Former OpenAI chief technology officer Mira Murati founded the lab in 2025, and it is based in San Francisco.

Company Size

201-500

Company Stage

Series A

Total Funding

$3B

Headquarters

San Francisco, California

Founded

2024

Get referred to Thinking Machines Lab

See people who can refer or advise you

Simplify's Take

What believers are saying

Equity is gaining value: investor Accel discussed a $40 billion-plus valuation in September 2026, versus $12 billion in 2025.
Researchers get big-lab computing power: in March 2026, Nvidia invested and committed at least a gigawatt of next-generation systems.
Fast hiring means room to grow: headcount rose from around 140 in April 2026 to roughly 200 by July.

What critics are saying

Leadership churn is high: four co-founders left between October 2025 and July 2026, the latest citing unsustainable stress and workload.
One product carries the business: its model, Inkling, is free, so revenue rests on Tinker, its paid tuning service.
Free and paid rivals beat its model: the company called Inkling, released July 2026, "not the strongest overall model".

What makes Thinking Machines Lab unique

Unlike research-only labs, it already has paying customers: revenue pace topped $100 million a year by September 2026.
Its next AI listens while it talks, like a phone call: a May 2026 preview replied in 0.4 seconds.
Its engineering has open-source roots: CTO Soumith Chintala co-founded PyTorch, the free software underpinning most AI research.

Help us improve and share your feedback! Did you find this helpful?

Benefits

Health Insurance
Dental Insurance
Vision Insurance
Unlimited Paid Time Off
Parental Leave
Relocation Assistance

Growth & Insights and Company News

Headcount

6 month growth

26%

1 year growth

24%

2 year growth

-3%

Africa.com Media Group

Oct 10th, 2026

Zimbabwean founders raise $2m for Ocular AI, selling voice data to the frontier labs.

Zimbabwean founders raise $2m for Ocular AI, selling voice data to the frontier labs. 10 October 2026 3 Mins Read Zimbabwean entrepreneurs Michael Moyo and Louis Murerwa have raised $2 million in pre-seed funding for Ocular AI - a company that already reports seven-figure revenue and counts frontier AI labs and Fortune 100 companies among its customers. Drive Capital led the round, with participation from Y Combinator, Alumni Ventures, 1745 Ventures (formerly Bertelsmann Digital Media Investments), Orange Collective, MyAsia VC and angel investors. A different position in the value chain. Most African AI companies in this space build applications for African markets. Ocular AI sells into the companies building the models. Founded in 2024, it produces datasets, evaluations and benchmarks used to train and assess advanced AI systems, with a focus on voice and audiovisual technology. Several top frontier AI labs and Fortune 100 companies now use its material, according to the company. That places two Zimbabwean founders upstream of the model developers rather than downstream of them - supplying inputs to the labs whose outputs most of the continent consumes. The technical argument. Traditional voice AI converts speech into text before generating a response, discarding tone, pauses, emphasis and timing along the way. Voice-native models process speech directly - listening and speaking simultaneously, responding to interruptions and shifts in a conversation. NVIDIA, Thinking Machines Lab, OpenAI and Google have all been developing or releasing audio-native models, which has raised demand for data that reflects how people actually talk. Ocular AI captures conversations at studio quality and builds evaluations designed to find where models fail. It has flagged the limited availability of high-quality full-duplex speech data, including continued reliance on older telephone-based datasets - a specific and checkable gap. The company's position is that the constraint is no longer data volume but data quality: material covering pauses, interruptions, overlapping speech and natural response, which is scarce in publicly available online sources. It is now extending into audiovisual, producing high-fidelity datasets for models that process voice, facial expression and gesture together in real time. Expertise as infrastructure. Ocular AI has built an Expert Network of thousands of vetted specialists who supply the human knowledge behind its datasets. Its method pairs domain experts with research: experts establish what a correct response looks like, research identifies where models fail, and the findings become datasets and evaluations. That is the same argument running through the benchmark problem generally - that evaluating a model requires knowing what good looks like, and that judgment has to come from somewhere. What hasn't been disclosed. No customers named, no revenue figure beyond "seven-figure," no Expert Network headcount, and no detail on where the company or its expert network is based. By Rugare Mubika, Next Africa.

Yahoo Finance

Sep 23rd, 2026

Thinking Machines Lab signs $65M deal with Crusoe for AI inference infrastructure

Crusoe, a vertically integrated AI infrastructure provider, has signed a $65 million annual agreement with Thinking Machines Lab to power inference for the lab's open models on Crusoe Cloud. Through Crusoe Managed Inference, Thinking Machines Lab will serve production workloads including its Inkling models, GLM 5.2 and 5.3, on NVIDIA HGX B200 systems connected with NVIDIA Quantum-2 InfiniBand networking. The service enables Thinking Machines Lab to scale without managing its own inference stack. The companies plan to expand into batch inference for large-scale synthetic data generation. Thinking Machines Lab joins other clients that have taken Crusoe Managed Inference past $100 million in contracted annual recurring revenue less than a year from launch.

GlobeNewswire

Sep 23rd, 2026

Crusoe and Thinking Machines Lab partner to power open-model inference at scale.

Crusoe and Thinking Machines Lab partner to power open-model inference at scale. The AI lab will run its production workloads on Crusoe Managed Inference, tuned for throughput, price-performance, and reliability. September 23, 2026 11:07 ET | Source: Crusoe SAN FRANCISCO, Sept. 23, 2026 (GLOBE NEWSWIRE) - Crusoe, the industry's first vertically integrated AI infrastructure provider, and AI innovator Thinking Machines Lab today announced a $65 million annual agreement to power inference for the lab's open models on Crusoe Cloud. Through Crusoe Managed Inference, Thinking Machines Lab will serve a range of production workloads - including its Inkling models, GLM 5.2 and 5.3, and its own fine-tuned variants - on a dedicated Tailored Deployment of NVIDIA HGX B200 systems connected with NVIDIA Quantum-2 InfiniBand networking, engineered for high-throughput, cost-efficient serving at scale. "Crusoe Managed Inference took us from evaluation to production quickly and met the standards we hold our own systems to, giving us the scale and economics to reinvest in the research at the core of what we do," said Myle Ott, ML Infra Lead at Thinking Machines Lab. "Thinking Machines Lab has a very sophisticated engineering team that expects partners to meet their high bar as they grow," said Erwan Menard, SVP, Product Management of Crusoe Cloud. "Our team of leading AI engineers built Crusoe Managed Inference to deliver cost-efficient, reliable infrastructure, inference, and model lifecycle tooling as-a-service, so innovative companies can run their favorite models in production at scale, relying on our expertise." Crusoe will run and support the cluster directly as a Tailored Deployment - a dedicated, benchmarked, SLA-backed endpoint - giving Thinking Machines Lab the price-performance and headroom to scale without managing its own inference stack. The companies are also looking to expand into batch inference for large-scale synthetic data generation, turning inference into a flywheel for research. Thinking Machines Lab joins a roster that has taken Crusoe Managed Inference past $100 million in contracted ARR less than a year from launch. About Crusoe As the AI factory company, Crusoe is on a mission to accelerate the abundance of energy and intelligence. The company provides a reliable, scalable, cost-effective, energy-first solution for AI infrastructure. By harnessing large-scale energy sources, building AI-optimized data centers, and delivering a powerful AI cloud platform, Crusoe empowers its customers and partners to build the future faster. For more information, visit crusoe.ai.

Toscale

Sep 5th, 2026

Thinking Machines Lab seeks $40B valuation as Nvidia extends its capital allocator role into model-space.

Thinking Machines Lab seeks $40B valuation as Nvidia extends its capital allocator role into model-space. Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, is in talks to raise between $5 billion and $6 billion at a pre-money valuation of at least $40 billion. Nvidia is separately in discussions to invest approximately $2.5 billion, while existing backer Accel is reportedly in talks to lead the round. The valuation represents a significant correction from the company's previous fundraise ambitions. Thinking Machines had sought a $50 billion to $60 billion valuation in a round that collapsed earlier this year, with three co-founders returning to OpenAI. The current talks price the lab at 3.3 times its July 2025 seed round, which raised $2 billion at a $12 billion valuation led by Andreessen Horowitz with participation from Nvidia, AMD, Cisco, ServiceNow, Accel, and Jane Street. The Nvidia dimension is what makes this more than a fundraise story. The chipmaker's reported $2.5 billion investment in Thinking Machines completes a five-week pattern of model-space capital deployment that has no historical precedent from a hardware company. Nvidia's reported $12.9 billion acquisition of Hugging Face (Aug 26-27) targeted model distribution. Its investment in SSI (Jul 28) backed safety-aligned frontier research. Its $30 billion-plus investment in Perplexity (Aug 24) bet on AI-native search. And now the Thinking Machines talks target open-weights frontier infrastructure. Each investment occupies a different layer of the emerging AI stack. Together they suggest Nvidia is no longer content to sell compute to frontier labs. It is becoming the capital allocator that determines which of them survive and on what terms. The partnership infrastructure reinforces the pattern. In March 2026, Thinking Machines and Nvidia announced a multiyear arrangement under which the lab will deploy at least one gigawatt of Nvidia's next-generation Vera Rubin systems, with first rollout targeted for early 2027. Capital investment and hardware supply are no longer separate conversations. For Thinking Machines, the round arrives at a critical juncture. The lab shipped Inkling-Small on July 15 - a 276 billion-parameter open-weights model with 12 billion active parameters that scored 80.2 percent on SWE-Bench Verified and 64.7 percent on Terminal Bench 2.1. The Tinker fine-tuning-as-a-service platform launched alongside it, positioning the lab to generate recurring revenue from enterprise customization rather than relying on inference fees alone. The open-weights positioning carries strategic weight in the current market. Thinking Machines is the only US-based lab shipping frontier-class open-weights models that compete directly with Chinese releases from DeepSeek, Moonshot, Qwen, and GLM on both capability and price. For investors weighing geopolitical risk and supply-chain provenance, that distinction commands a premium - though it has not yet translated to the valuations the company once sought. The repricing from $50-60 billion to "at least" $40 billion tells its own story. The failed round earlier this year was not just a financing miss. The departure of three co-founders back to OpenAI signaled organizational turbulence that makes this second attempt a credibility test as much as a capital raise. Whether the lower valuation reflects broader market cooling, company-specific concerns, or both, the number itself becomes a signal for how the market is pricing frontier-lab risk in late 2026. The deal has not closed. All reports describe ongoing discussions, and the final structure - whether Nvidia's $2.5 billion is part of the $5-6 billion total or sits alongside it - remains unclear. Accel and Thinking Machines have not commented. What to watch: whether the round closes at or above the reported $40 billion floor, and how Nvidia's dual role as hardware supplier and equity investor shapes Thinking Machines' deployment timeline for the Vera Rubin systems commitment.

MarketScreener

Sep 3rd, 2026

Thinking Machines Lab aims to raise $1 billion at $40 billion valuation, the Information reports.

Thinking Machines Lab aims to raise $1 billion at $40 billion valuation, the Information reports. Published on 09/03/2026 at 03:08 pm EDT or by email NVIDIA Corporation is the world leader in the design, development, and marketing of programmable graphics processors. The group also develops associated software. Net sales break down by family of products as follows: - computing and networking solutions (89%): data center platforms and infrastructure, Ethernet interconnect solutions, high-performance computing solutions, platforms and solutions for autonomous and intelligent vehicles, solutions for enterprise artificial intelligence infrastructure, crypto-currency mining processors, embedded computer boards for robotics, teaching, learning and artificial intelligence development, etc.; - graphics processors (11%): for PCs, game consoles, video game streaming platforms, workstations, etc. (GeForce, NVIDIA RTX, Quadro brands, etc.). The group also offers laptops, desktops, gaming computers, computer peripherals (monitors, mice, joysticks, remote controls, etc.), software for visual and virtual computing, platforms for automotive infotainment systems and cloud collaboration platforms. Net sales break down by industry between data storage (88.3%), gaming (8.7%), professional visualization (1.4%), automotive (1.3%) and other (0.3%). Net sales are distributed geographically as follows: the United States (46.9%), Singapore (18.2%), Taiwan (15.8%), China and Hong Kong (13.1%) and other (6%). Mean consensus Number of Analysts Last Close Price 230.36 USD Average target price 327.13 USD Spread / Average Target

Thinking Machines Lab

Developer Relations Engineer

About the job