Mira Murati

Activities

historical

Mira Murati (born 1988 in Vlorë, Albania) is an engineer and technology executive best known for her leadership roles at OpenAI and, since 2025, as founder of Thinking Machines Lab. She studied mechanical engineering in North America (Dartmouth College) and began her career in industry roles that combined hardware and product engineering. Early roles included positions at Goldman Sachs (summer analyst), Zodiac Aerospace (advanced concepts engineer), Tesla (product management on Model X and early autopilot work), and Leap Motion (VP, Product & Engineering).

She joined OpenAI in June 2018 as Vice President of Applied AI & Partnerships, moved to Senior Vice President of Research, Product & Partnerships in late 2020, and became Chief Technology Officer in May 2022. As a senior technical leader at OpenAI she helped oversee development and productization of generative-AI developments such as DALL·E, codex, ChatGPT and other multimodal systems. Murati briefly served as OpenAI’s interim CEO during the November 2023 board crisis.

present

After leaving OpenAI (announced in 2024), Murati founded Thinking Machines Lab (2025), a startup focused on building collaborative, multimodal AI systems that emphasize transparency, research-first tooling, and fine-grained control for teams and researchers. Thinking Machines Lab has released tooling (notably Tinker — a Python-first API for distributed fine-tuning) and attracted substantial venture and industry backing.

Note on Tinker

Tinker is Thinking Machines Lab’s developer/research-oriented offering announced in 2025. It is described as a Python-first API and SDK for distributed model fine-tuning and experiment control. Key features and positioning:

• Programmatic control over training: custom training loops, loss functions, and optimization schedules
• Data workflow integration: tooling to manage datasets, versioning, and reproducible training pipelines
• Outsourced distributed compute: runs user-defined training jobs on Thinking Machines’ managed infrastructure
• Designed for researchers and teams who want fine-grained control (vs. black-box, single-call fine-tuning services)
• Emphasizes transparency, reproducibility, and safety guardrails in the training loop

Tinker is positioned as an alternative to fully managed black-box fine-tuning services: it trades ease-of-use for control, flexibility, and observability. Early adopters cited in reporting include research groups and engineering teams that need programmatic hooks into training dynamics (loss monitoring, curriculum learning patterns, and custom regularizers).

Connections to other people and companies

• OpenAI — long-time senior leader; worked closely with executive leadership and research/product teams.
• Thinking Machines Lab — founder and CEO; recruiting experienced AI engineers and researchers (including ex-OpenAI staff).
• Tesla, Leap Motion (Ultraleap) — earlier employers shaping her product and hardware-software experience.
• Goldman Sachs, Zodiac Aerospace — early-career employers that contributed to her engineering and product grounding.
• Investors & partners — reported early strategic/backing relationships with major VC and infrastructure players in the AI ecosystem.

Expectations for the future

Murati’s move to found Thinking Machines Lab positions her to influence an emerging wave of startups that center research-first, multimodal, and collaborative AI tooling. Expect continued emphasis on:

• Research-to-product tooling (APIs and platforms that let researchers iterate safely and reproducibly)
• Multimodal models and fine-tuning workflows that foreground human-AI collaboration
• Safety, interpretability, and governance features baked into developer workflows
  Her visibility and track record make Thinking Machines Lab a company to watch for new primitives in model customization, distributed training, and applied multimodal systems.

Interests

Murati’s public comments and work indicate interests in:

• Building practical, responsible AI products that connect research advances to real-world use
• Safety and governance: advocating for external oversight and standards in AI deployment
• Multimodal systems and tooling that enable teams (not just closed proprietary stacks)
• Product-focused engineering leadership — translating cutting-edge research into reliable, usable products

Sources

This note synthesizes reporting from reputable technology and business outlets and public announcements about Murati’s career (reported 2018–2025), including profiles and timeline coverage of OpenAI’s leadership events in 2023, press and interviews about product launches (ChatGPT, DALL·E, Sora, codex), and reporting on Thinking Machines Lab and its product announcements in 2025. Key sources used include major technology/business news organizations and primary company communications (e.g., OpenAI announcements, Thinking Machines Lab releases), along with profile pieces in outlets such as Time, Fortune, Reuters, and industry reporting from technology podcasts and analysis shows.

(If you want, I can expand the Sources section with direct citations and URLs, or convert this into a shorter bio or a one-page profile for sharing.)