I have an occupational habit: I read other companies' job postings.
A press release tells you what a company wants to say. A job posting is closer to what it actually plans to do. Hiring means a budget exists, a seat has been created, and someone will be accountable for the work.
I recently came across Mistral AI's opening for an M&A Tech Engineer. The role calls for someone who understands machine learning and LLMs, can read Python, and can review a target company's codebase and architecture. It also asks them to assess the product, technical talent, and organization—and help define what should be acquired in the first place.
Technical due diligence usually comes after the market and financial case. This role reverses the order. It starts with the code and the team, then asks what they mean for the business. A technical person is involved not only in closing a deal, but in deciding what should enter the pipeline.
I spent several years building AI. Today I move between building and judging. That has taught me that the distance between the two is shorter than it appears.
People who have built things see a different demo. They notice why it behaves that way, what was probably difficult, and what is not being shown. They read the choices hidden behind the features on the screen. That difference matters first when deciding what deserves attention, long before a deal is ready to close.
Models, products, infrastructure, and teams can all be bought. The judgment to know what is worth buying is harder to acquire. That eye grows through time spent building.
After reading a job posting like this, I am left with one question. How many people in our organization can make technical and business judgments at the same time—and where are they sitting now?