The useful clue in Sasha Lascar’s career is a phrase borrowed from French kitchens: mise en place. Everything in its place. Before service begins, knives are sharpened, ingredients measured, stations arranged. The point is not tidiness for its own sake. The point is to make calm possible when the orders start arriving faster than anyone can politely discuss them.
Lascar learned the phrase in earnest during the summer of 2024, when he traded Stanford’s machine shops for the pastry section at Restaurant Aan de Poel in the Netherlands. The restaurant held two Michelin stars. The engineering graduate arrived as an intern. By his account, he was trusted to run an entire shift in his third week. His explanation was neither luck nor culinary genius. He asked questions, prepared obsessively and learned which details mattered at which moment.
Two years later, his materials have changed again. He is the founding AI Ops operator at Laurel, the enterprise software company that applies artificial intelligence to the stubborn business of recording how professional work gets done. The station is now a company. The ingredients are Slack messages, customer knowledge, workflows and decisions. The rush never quite ends.
A maker changes materials
Long before he built agents, Lascar built things that could bruise a toe. At Stanford he studied mechanical engineering, first for a bachelor’s degree, earned with distinction, and then for a master’s. His training ran through the Product Realization Lab, the campus workshop where ideas must eventually confront metal, plastic, tolerances and the ancient hostility of objects that refuse to fit.
As a senior, he described a practical vocabulary that included the CNC mill, manual mill, lathe, laser cutter and 3D printer. He could weld, solder and work in CAD. The list reads like a hardware store with entrance exams. Later, he became a course assistant in the lab, a role selected through a competitive process and built around helping other students prototype, experiment and discover why the first version was optimistic.
His favorite projects were charmingly specific: ravioli molds, rolling pins and cake molds made for his own kitchen. Even the hobbies carried the same signature. The finished object was useful, but the attraction lay in designing the process that made it repeatable. A mold is a tiny argument for leverage.
A senior capstone project widened the lens. Lascar and four teammates designed “Combo,” a countertop system intended to make food-waste collection feel less like a household penance and more like the beginning of a garden. Their need-finding suggested that people disliked composting but liked gardening. The team joined the two experiences. It was a modest product with a durable lesson: behavior changes when an awkward handoff becomes an appealing loop.
The long apprenticeship in operations
Stanford supplied more than machine tools. Lascar moved through the student institutions that orbit Silicon Valley’s startup economy. He worked on TreeHacks, served in multiple roles at the Business Association of Stanford Entrepreneurial Students and eventually became its co-president. He held a venture fellowship with Xfund, worked in program operations at Floodgate and later joined the General Catalyst Venture Fellows and Rough Draft Ventures network.
This could look like restless résumé collecting. The sequence makes more sense as an education in organizations. A machine has visible parts and known interfaces. A company contains people, incentives, half-remembered decisions and spreadsheets called “final_v7.” Both fail at their connections.
The pastry internship was the apparent detour that revealed the pattern. Lascar wrote about shoshin, the beginner’s mind, and about kaizen, the practice of continuous improvement. He treated the top of a plated dessert as a north star. He learned to stay busy while the kitchen was calm so he could remain calm when it was busy. A Michelin kitchen compressed systems thinking into something immediate: a missing component could not be patched in next quarter’s release.
He also discovered a humane constraint. Excellence, he wrote, did not require being a jerk. The sentence sounds almost comic because so many high-performance cultures pretend otherwise. But it matters to the work that followed. Operations is not only the removal of waste. It is the design of conditions in which people can do ambitious work without inventing fresh chaos for their colleagues.
One hundred agents and the problem after the demo
At Laurel, Lascar’s remit crosses product, sales, marketing, operations and finance. In February 2026, after running AI Ops as a team of one for six months, he said more than 100 agents had been deployed across the organization. The company then began recruiting specialized AI Ops roles for product operations, go-to-market, finance and a generalist track.
A hundred is the sort of number that behaves well in a slide deck. Lascar’s more interesting observation concerns what happens next. Give every employee access to AI and they will build. One person makes a note vault. Another makes an agent for call preparation. A third solves the same problem in a different corner. From inside each team, progress is obvious. From above, the company may resemble a suburb where every house has constructed its own power station.
Lascar argues that personal context is the easy portion. The difficult portion is getting context to travel. A chief-of-staff agent can draft an email or summarize a meeting, but useful work rarely belongs to one person alone. Inputs arrive from other teams. Outputs become someone else’s inputs. Customers, projects and commitments live across shared surfaces. An assistant trained only on one employee’s world has, in Lascar’s memorable description, merely skimmed the job description.
The multiplayer company
This is where the maker’s instinct returns. The task is not to collect the most tools. It is to shape interfaces so that a discovery in sales can help product, a finance constraint can reach operations and a customer insight can survive beyond the person who first heard it. Lascar describes the desired result in compounding terms: what one team figures out, the next team inherits.
The company as a prepared station
The aspiration is explicit. Lascar wants Laurel to become the most productive company per headcount by reimagining how it operates in the age of AI. Productivity claims deserve suspicion; they are often a euphemism wearing an expensive fleece. His framing, though, emphasizes removing duplicated effort and returning people to what he calls their zone of genius.
That ambition also explains the people he says he wants around him: builder-operators, future founders and colleagues who notice a broken process and feel compelled to repair it. It is a recruiting pitch, certainly, but also a description of the culture his systems require. Shared infrastructure only compounds when people contribute to it. An immaculate private agent may save its owner an hour; a plain tool adopted across five departments may alter the company’s week. The operator’s taste must therefore extend beyond invention to adoption, maintenance and the delicate politics of asking busy people to change how they work.
There is a neat fit with Laurel’s external business. The company sells technology meant to recover time that lawyers, accountants and other professionals fail to record, then turn work activity into clearer decisions about capacity, billing and profitability. Lascar works on the inward-facing version of the same question: where does effort disappear inside the organization, and how can the useful trace of it remain?
His background makes him a revealing figure in the emerging practice of AI operations. He is not presented as a research scientist chasing a benchmark. He is an operator who can read a workflow, an engineer comfortable with prototypes and a former pastry intern who knows that elegance during service is purchased by unglamorous preparation beforehand.
The career still has the pleasing uncertainty of an early prototype. Lascar’s public record runs from volunteer construction work and student venture groups to food robotics, composting, kitchen tools and enterprise agents. The jumps feel large only if industries are the unit of measurement. Use systems instead, and the path straightens.
A mold gives shape to something repeatable. A kitchen station puts the next action within reach. Shared context lets the next team begin where the last one stopped. Each is a way of arranging the knowns before the unknowns arrive. The material changes. The workbench remains.