Shruti Gandhi writes the first check, teaches the CS class, and still files herself under engineer

Read Shruti Gandhi's bio quickly and you will see a venture capitalist. Read it as written, which is 'VC Eng,' and you get the actual person. The engineering is not a decorative origin story that ended when she got a fund. She spent her early career building: mainframe security at IBM, collaboration tools on Lotus Sametime, then an internal incubator and a think-and-do tank that reported into the chairman's office. Then True Ventures and Samsung's venture arm, and then her own fund.

The proof that the technical half persists is that she still teaches computer science at Columbia while running Array Ventures. Most investors who say they are technical mean they read the docs. She grades the homework.

Solo, pre-seed, and actually returning

Array is a pre-seed fund writing first institutional checks in the $250K to $2M range, aimed at technical founders building enterprise data, AI infrastructure, security, and developer tools. That is the hardest end of the market to be right in, because at pre-seed there is no traction to analyze and often no product, so the only real input is whether you can evaluate the person and the technical bet.

Fifteen realized exits from pre-seed positions, as a solo general partner, is the part that should make you sit up. Solo GPs are common now; solo GPs with a long list of actual liquidity are not. The portfolio reads like a list of things that became infrastructure: Drata, Placer.ai, Pendo, HappyRobot.

And the thinking is public. She publishes her theses, sometimes literally as a handwritten list of what she wants to fund, everything from AI governance infrastructure and edge intelligence to embodied world models, ending with the entry that gives the whole thing away: a boring email company, just make my inbox better. A serious technical taxonomy with a joke at the bottom is a good sign about a person.

The teaching is the tell

There is an easy version of her career: take the exits, stop building, stop teaching, do panels. She kept the classroom. Teaching CS while running a fund is a strange allocation of a very expensive calendar unless you actually like the material, and liking the material is more or less the whole qualification for backing people at pre-seed, where there is nothing else to go on.

If you are a technical founder with no traction and a hard problem, her substack and her theses are the cheapest way to find out whether you are building something she would fund before you ever email her.