I thought I knew how to manage a team. I’d done it for years, built companies from the ground up, and even had a couple of successful exits. RemoteTeam, which was acquired by Gusto, was my baby. We were remote-first before it was cool. So when I started building a new AI venture, I figured managing the team remotely would be the easy part. I was wrong. Terribly wrong.
Managing a team of brilliant AI engineers, data scientists, and researchers scattered across the globe is a different beast entirely. The usual management playbook? Toss it out the window. The things that worked for my past teams—even the remote ones—failed spectacularly. For the first six months, I felt like I was failing. We missed deadlines. Communication was a mess. The team was brilliant, but they weren't a team. They were just a collection of smart people working in silos.
After a lot of sleepless nights, and more trial and error than I care to admit, I finally figured out what works. It wasn't what I expected. The most effective strategies were often the most counterintuitive. Here are the five hard-won lessons that turned things around, saved my sanity, and ultimately, made our remote AI team thrive.
1. Stop Trying to Replicate the Office
My first instinct was to recreate the office environment online. I scheduled daily stand-ups, set up a dozen Slack channels for 'watercooler' chat, and even tried virtual happy hours. It was a disaster. The team hated it. The stand-ups felt like a waste of time for people working on complex, deep-thinking problems. The social channels were a ghost town. It felt forced because it was.
Here’s the thing about top AI talent: they need long, uninterrupted blocks of time to do their best work. My constant attempts to foster 'togetherness' were actually just interruptions. The big realization was that remote work isn't about replicating the office; it's about creating a new, better way of working.
So, I killed the daily stand-ups. We moved to a weekly, more in-depth sync where we discussed progress and roadblocks on a higher level. I replaced the forced social events with an optional monthly 'demo day' where people could show off what they were working on, no matter how small. The focus shifted from looking busy to being productive. The result? A 30% jump in completed experiments and a team that was genuinely more connected because they were bonding over the work itself, not forced small talk.
2. Hire for Writing, Not Just for Coding
This one might sound strange. Why would I prioritize writing skills for a team of engineers? Because in a remote-first, asynchronous environment, writing is thinking. Clear writing is clear thinking. I can’t walk over to someone's desk to get a quick update. Everything—from project proposals to bug reports to code reviews—has to be communicated through text.
I once had a machine learning engineer who was a genius. His models were incredible. But his documentation was a nightmare. His code comments were non-existent. Getting information from him was like pulling teeth. It created a massive bottleneck. No one else on the team could build on his work without spending days trying to decipher it. He was a brilliant individual contributor, but a terrible remote team member.
Now, a key part of my interview process for any technical role is a writing test. I give candidates a complex technical problem and ask them to write a short proposal for how they would solve it. I'm not looking for a perfect essay. I'm looking for clarity, structure, and the ability to explain a complex idea simply. The candidates who can write well are almost always the ones who can think clearly, collaborate effectively, and thrive in a remote setting. It's a surprisingly effective filter.
3. Your Best People Don't Need 'Managing'
I used to be a very hands-on manager. I liked to be in the details, to know what everyone was working on at all times. With a remote AI team, this is not just inefficient; it's insulting. You're hiring people who are at the top of their field. They are experts. The last thing they need is someone micromanaging their every move.
My role shifted from being a manager to being a facilitator. My job isn't to tell them how to do their work. My job is to give them a clear goal, make sure they have the resources they need, and then get out of their way. I protect them from distractions. I run interference with other departments. I make sure they have the freedom to explore, to experiment, and even to fail.
One of the best engineers I ever hired told me in his first week, "Just point me at the mountain you want me to climb. Don't tell me which path to take." That stuck with me. I now define success by outcomes, not by process. I don't care if someone works from 9 to 5 or from 9 pm to 5 am. I don't care if they take a three-hour break in the middle of the day. What I care about is this: are they solving the problem? Are they delivering results? This level of autonomy is scary at first, but it's the only way to attract and retain the kind of talent that can build the future.
4. Over-Invest in Your Onboarding
When you're remote, you can't just have a new hire shadow someone for a week. The onboarding process has to be incredibly deliberate and structured. A poor onboarding experience is the fastest way to lose a new hire, especially in a competitive field like AI. I learned this the hard way. We lost a fantastic data scientist in her first month because we just threw her into the deep end. She was overwhelmed, confused, and felt disconnected from the team. It was my fault.
Now, our onboarding is a four-week-long, meticulously planned program. It includes:
- A dedicated onboarding buddy: Not their manager, but a peer who can answer all the 'stupid' questions.
- A structured learning path: A checklist of documentation to read, people to meet, and starter projects to complete.
- Early wins: We design a small but meaningful first project that they can complete and ship within their first two weeks. This builds confidence and makes them feel like a contributing member of the team right away.
- Forced social interaction (the good kind): We schedule a series of 1-on-1s with every member of the team in their first month. No agenda, just a 30-minute chat to get to know each other as people.
It's a huge investment of time and resources. But the payoff is massive. Our retention has gone up significantly, and new hires are fully productive in half the time they used to be. In a remote world, you have to build the connections that would happen organically in an office.
5. The Metrics You Track Determine the Culture You Build
In the early days, I was tracking the wrong things. I was looking at lines of code, hours worked, and the number of tasks completed. These are vanity metrics. They tell you nothing about what actually matters: impact.
An engineer could write a thousand lines of code that do nothing to improve the model's accuracy. Another could write ten lines that result in a breakthrough. An AI team isn't a factory. You can't measure productivity by the number of widgets they produce.
We had to completely rethink how we measured success. We started focusing on outcome-based metrics that were directly tied to our business goals. For our research team, it was about the number of successful experiments that led to a measurable improvement in our core product. For our engineering team, it was about the speed and reliability of our model deployment. We created a culture of accountability around results.
This shift had a profound effect on the team. It clarified what was important. It gave them a clear sense of purpose. And it empowered them to make their own decisions about how to best achieve those goals. When you start measuring what matters, people start focusing on what matters.
The Remote Advantage
Managing a remote AI team is the hardest thing I've ever done. But it's also the most rewarding. When you get it right, you can build a team of world-class talent that is more focused, more productive, and more engaged than any in-office team I've ever led. You just have to be willing to throw out the old rules and embrace a new way of thinking. It's not about managing people; it's about creating an environment where brilliant people can do their best work, wherever they are in the world.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.