How to Answer: "Why Do You Want to Work at Thinking Machines Lab?"

Answer this question by naming the lab's real work and connecting your own work to it. Thinking Machines Lab is the AI research company started by Mira Murati, the former chief technology officer of OpenAI. Its first product is Tinker, an API (a programming interface) that lets researchers fine-tune open source models on managed GPU clusters. Fine-tuning means adjusting a trained model with your own data. A strong answer proves three things. You know this product. You want to build open, controllable AI tools. And you have evidence that you can contribute.

The team is small, so every hire changes the company. The interviewer uses your answer to test whether you understand what the company actually does. General enthusiasm about AI does not pass that test.

What the Interviewer Is Listening For

1. Specific knowledge of the lab. The company publishes real material. Tinker has public documentation. The research blog, called Connectionism, has posts such as "LoRA Without Regret" and "Defeating Nondeterminism in LLM Inference". LoRA is a method that trains a small add-on to a model instead of the whole model. Quoting one concrete detail proves you did more than read the news.
2. A reason for this lab and not another. Most large labs sell a chatbot to consumers. Thinking Machines Lab sells researchers control over model training. Your answer should show why that difference matters to you.
3. Evidence of ownership. The team is small, so each engineer covers a wide area. Interviewers listen for projects you completed from start to finish.
4. Depth in the relevant stack. The company's daily problems are training infrastructure, GPU scheduling, and model serving. Your motivation is more convincing when your skills match those problems.

A Three-Part Structure

Part 1: Your reason (2 to 3 sentences). Name the specific work you want to join. Include proof that you have used or studied it.

Part 2: Your evidence (3 to 4 sentences). Describe projects you built and owned. Use numbers: users, latency, cost, scale.

Part 3: Your direction (1 to 2 sentences). Say what you would work on in your first year. A specific direction shows you thought about the job, not just the brand.

Sample Answer

"I want to work here because Tinker solves a problem I have lived with. My team fine-tuned open models for two years, and we spent most of that time on infrastructure instead of research. I built our training pipeline on a shared GPU cluster, including checkpointing and failure recovery for jobs that ran for days. When Tinker launched, I read the documentation and ran a small LoRA training job that same week. I also follow the Connectionism blog, and the post on nondeterminism in inference described bugs I had spent weeks debugging. I want to build the platform layer that lets researchers spend their time on ideas instead of infrastructure. That is the work I do best, and this company does it in the open."

Every claim in this answer can be checked. It names the product, shows real use, includes owned work, and ends with a direction.

Common Mistakes

• A generic AI answer. If your answer also fits OpenAI or Anthropic word for word, it fails here. Name the product and the research.
• Praise without evidence. Saying the team is excellent is not a reason you fit the team. Connect your skills to their problems.
• Only naming the founders. Admiration for Mira Murati is not a contribution. Interviewers want to know what you will build.
• No contact with the product. The Tinker documentation is public. Arriving without having read it signals low interest.

How to Prepare

• Read the company's own material. Work through the Tinker documentation and two or three Connectionism posts. Bring one concrete detail from each into your answer.
• Learn the full loop first. Read What is the Thinking Machines Lab interview process like? so you know where this question appears.
• Prepare your evidence stories. Practice the questions in Top Thinking Machines Lab behavioral interview questions. The Grokking Modern Behavioral Interview course teaches the story structure.
• Prepare the technical rounds too. Review What to expect in the Thinking Machines Lab system design interview for the design topics that match the product.