Farhan Thawar leads engineering at Shopify — roughly 3,000 people building the infrastructure behind one of the largest commerce platforms in the world. He joined in 2019 when Shopify acquired Helpful.com, the company he co-founded, and he's been there now for longer than any job he's held before.
This conversation for the StartWell Podcast moves fast, starting with a rapid-fire round before digging into all sorts of great topics and Farhan gives us his take on AI vs. headcount, the most overrated skill in engineering right now, the one line he'd delete from every job posting.
From there, we get into how Shopify actually operates: why they build almost everything in-house, including their own alternative to Jira, why "infrastructure over features" is the default answer to nearly any product decision, how Sidekick, their AI tool, is changing what merchants can do without a developer... And why Farhan thinks most CTOs still have AI's job description wrong - treating it as a cost-cutter when it's really a ceiling-raiser.
We also cover how Shopify runs culture at scale: 1,000 engineering interns a year, a 7,000-person hackathon (the largest in the world) and a philosophy he calls "heightening the misalignment," which is really just a case for being upfront about what kind of company you are.
And underneath all of it is Farhan's own founder story: three pivoted products in three years at Helpful.com, the acquisition, and a still-active habit of angel investing across 106 companies, less for the returns than to stay close to the people building things.
Advice Farhan offers in this episode:
- Re-test AI tools on a schedule, not just once. If something failed six months ago, try it again — the gap between "not there yet" and "better than humans" can close in a single quarter.
- Choose infrastructure over the one-off feature, even when the feature is faster to ship. The infrastructure enables things you haven't thought of yet.
- Make your work culture legible, not comfortable. State plainly whether you're remote, hybrid, or IRL — ambiguity does more damage than an answer someone might dislike.
- Don't build what you can buy — unless you have an actual opinion about how it should work differently. Building for its own sake is waste; building because you disagree with the default is strategy.
- Protect focus by making it harder to start distracting work, not easier to finish it.
