You have 1 article left to read this month before you need to register a free LeadDev.com account.
Estimated reading time: 7 minutes
Key takeaways:
- At Spotify, a high-intensity engineering culture means more done per hour, not more hours.
- AI raises the stakes on review and deletion.
- Autonomy and psychological safety make the intensity sustainable.
When you hear the phrase “high-intensity culture,” you might imagine a stressful work environment where employees are sweating to meet tight deadlines, working overtime, and leadership is obsessed with productivity performance metrics. However, high-intensity doesn’t have to mean high-stress.
For Farhan Thawar, head of engineering at Shopify, creating a high-intensity engineering culture is less about demanding results, and more about enacting smart leadership choices that optimize and guide engineering talent toward agile ways of working.
“Intensity isn’t about working more hours,” says Thawar. “It’s about getting more done per hour. The goal is to make every hour count, not add more of them.”
For Shopify, making those hours count means reducing unnecessary meetings, time-boxing releases, encouraging pair programming, and entrusting employees with ownership of their work, including work produced with AI agents.
It’s a culture with enough psychological safety to ask dumb questions, retry ideas, and delete code. This sort of high-intensity culture can unlock time for deep focus, constant iteration, and even cost-savings.
Your inbox, upgraded.
Receive weekly engineering insights to level up your leadership approach.
Making the hours count
Parkinson’s Law states that work expands to fill the time available for its completion. In other words, jobs tend to fill whatever time they’re given. If you don’t put time constraints on a project, it’s likely to drag on and on.
“If you have a month, you’ll use a month,” says Thawar. “So we create urgency.” At Shopify, this is done by setting a default weekly feedback cadence.
“We ask the team to deliver shippable work every week,” he adds, “not because we’ve decided to lower the quality bar, but because we don’t think the tradeoff between quality and time is real.”
Rather than let work stagnate, the goal is to move fast and learn along the way. “Being wrong is less costly than being slow, if you’re good at course correcting,” explains Thawar. “Decide fast, fix fast.”
That velocity depends heavily on automated testing. “We run hundreds of thousands of automated tests before anything goes to production.”
Eliminating meetings
A quick development pace sounds great on paper, but there’s an obvious enemy of deep developer work: meetings.
Shopify gained media attention when, in 2023, it cancelled 322,000 hours worth of internal recurring meetings. “We called it Meeting Armageddon,” says Thawar. The massive calendar purge is a practice they still employ.
As Thawar explains: “Once a year, at a random time, we delete all internal recurring meetings. Then there’s a two-week moratorium where you’re not allowed to schedule a new recurring meeting. It forces you to actually think: do we need this meeting?”
By ruthlessly eliminating meetings that don’t belong, Shopify avoids unnecessary time-wasters – especially the regularly scheduled weekly meetings that are just “on the books.”
According to Thawar, this has freed up lots of uninterrupted time for deep engineering work. He’s also noticed other interesting positive side effects, such as improved asynchronous communication and more intentionality when booking meetings.
More like this
Adopting pair programming
Another core element of Shopify’s engineering culture is pair programming. This is when two engineers work together in real time, either physically side-by-side or remotely, on the same coding task.
Thawar describes the activity as Shopify’s “secret weapon.” Its engineers typically pair program for about eight to 16 hours a week, or a few hours here and there.
“Pair programming doesn’t just improve code quality, it forces intense collaboration, improves knowledge sharing, reduces distractions, and leads to faster unblocking,” he says. “When two people are in it together, you can’t get stuck for days.” That shared context is especially useful for debugging and quickly testing ideas.
Pair programming dates back to the company’s beginnings, when Shopify’s CEO and the original CTO would pair for an hour at a time on different tasks. “If the problem couldn’t be solved in an hour, they would delete all the code (not the tests!) and start over,” says Thawar. “Some of that original pair-programmed code is still live in the code base today.”
Owning the output of AI
Like most fast-moving software engineering organizations, Shopify developers are making use of AI agents in their daily work for scaffolding, coding, testing, and bug-fixing.
According to Thawar, there’s nothing wrong with leaning on large language models (LLMs) to generate 90-95% of your code. What matters, however, is understanding every line of generated code. “They [developers] still need to be able to look at a line and say ‘that’s wrong,’ and fix it themselves,” he says. “The comprehension bar doesn’t move.”
Thawar says this could change as the industry develops more robust AI harnesses that meet Shopify’s internal quality standards. For now, the onus is still on developers to own what they create, whether AI is involved or not.
“Our approach is simple: pair with an AI agent, but your name is on the pull request (PR),” he says. “You have to understand it before you submit it.”
Questioning, deleting, and rewriting
AI dramatically lowers the bar for generating code. A single prompt can generate thousands of lines of code, potentially burdening reviewers and bloating the codebase. That makes review and deletion even more important.
As Thawar says: “A smaller codebase is easier to understand, less error-prone, and faster to navigate. That’s always been true. The AI era makes this more urgent.” Constant pruning means less infrastructure to maintain, reduced latency, and simpler code to understand, he says.
Code deletion is a big part of the Shopify culture. This is very apparent at Shopify’s Hack Days, an internal hackathon it runs twice a year, explains Thawar.
“One of my favorite teams that shows up every time is the Delete Code Club. Their entire purpose is to delete code from Shopify. At any Hack Day, we routinely remove over a million lines. In one calendar year, we deleted over five million.”
This mindset requires a level of comfort where people can question prior work. “The other thing we focus on is psychological safety, creating a culture where it’s genuinely okay to ask ‘stupid’ questions,” says Thawar. “A daily willingness to fail in public is, I’d argue, a superpower.”
So, they’ve intentionally fostered a culture where asking basic, constructive questions is encouraged. This could be as simple as:
- Why does this code exist?
- Why does this meeting exist?
- Why does this process exist?
“Most of the time, nobody remembers,” says Thawar. “If nobody remembers, you have your answer.” Sometimes, that willingness to question prior work can extend to entire rewrites.
New York • September 15 & 16, 2026
Don’t figure it out alone. Learn what works at LDX3 New York.
Trusting the team
A final core aspect of Shopify’s culture is autonomy. Developers tend to perform well when there’s a clear mission and authority to execute on it. “You give them the mission, you remove toil, and you trust them to figure out how they can help the customer,” says Thawar. “People do their best work when they own the outcome.”
This isn’t limited to full-time employees, either. Thawar says that engineers learn a lot from interns, who make up about 10% of the engineering organization. “They show us what’s new, what’s possible, and they curiously always ask why we do things the way we do. They’re a secret weapon.”
In fact, that secret weapon has delivered tangible results. Thawar recalls a moment when one intern deleted six lines of code, resulting in incredible cost savings. “That single change saved us over $600,000 a year in infrastructure costs.”
Customer empathy also helps sustain a sense of urgency. In Shopify’s case, this means identifying with the end users, who tend to be entrepreneurs building their business at a quick pace. “Sharing this urgency allows us to be intense and show up for our customers,” he says.
Managers need to shift, too
Lastly, Thawar recommends being part of the team you’re leading. “Today, managers are making deep technical decisions, they are coding (along with their AI agents), and they are closer to the real work than in the past.”
Directors don’t have an excuse now. “In the past, it was possible to not be in the details of your team’s work, or get away with not contributing code. Now with AI, it’s impossible to not be in the details of building.”
Staying this close to the work helps leaders see what’s actually possible with AI tools and make more informed tradeoffs based on experience.
Altogether, these practices create a fast-moving engineering organization designed to support the business. “We really do care what the inside of Shopify looks like, even if our customers never see it,” he adds.
