How to Approach Remote Team Communication

In 2013, a small team in San Francisco set out to help people organize their ever-growing number of digital photos and memories. After securing billions of photos and videos from tens of millions of users in 95 countries, they came to realize the true value of their work.

18 months ago, under the leadership of serial entrepreneur Doug Aley, the team decided to productize their face recognition technology. The company, now known as Ever AI, offers a best-in-class technology with the most comprehensive, real-life dataset of any private company in the world. Their models have identified hundreds of millions of clustered identities to date, and they are changing how the world approaches security, marketing personalization, and the marriage of online and offline experiences.

Ever AI’s growth has been powered by the expansion and productivity of its remote engineering team. The machine learning power needed to bring a service like Ever AI to life is immense. Like many Bay Area startups, the company knew how difficult it is to recruit locally and decided to go the remote team route.

The struggle to recruit top talent locally is well-documented. As data science, the field in which machine learning belongs, has become increasingly important across all industries, the talent gap in San Francisco has become unwieldy. There are over 31,000 more roles requiring the skillset than there are qualified workers. Demand for machine learning engineers is the fastest growing subset of the data science group. To secure the right expertise and experience, going remote isn’t just one way for startups to grow machine learning teams in the Bay Area, it’s the only way.

With the help of Terminal, Ever AI built teams in three cities; Vancouver, Kitchener-Waterloo, and Toronto. They now have 12 engineers, and according to Chief of Staff Tiffany Lee, the quality of those hires exceeded all of their expectations. “What we’ve found in Canada are talented, hard-working engineers who really put what we are trying to achieve as a company first — it’s an incredible group.”

Although Tiffany and her team have reaped the benefits of a global talent pool, they admit there’s plenty of work to be done to improve how their team operates. First on the list is team communication, a fact not lost on Ever AI. The team has worked hard to develop efficient systems and learned some hard lessons on the road to where they are now.

Here are five insights that every company that is contemplating remote teams should consider:

Nope, Blocking Time On Your Calendar Isn’t Good Enough.

Distributed employees will often tell you they struggle with access to leadership, which causes them to feel more like contractors than full-time team members. Leadership teams that are conscious of this fact often opt to block time on their calendar just like university professors reserve time for office hours. In theory, this practice ensures remote team members access to everyone, but in reality, it causes anxiety.

“Great communication can’t be forced — leadership teams need to be just as accessible for a remote team member as the employee that is physically sitting right next to them,” says Tiffany. The solution, according to both Tiffany and Ever AI’s Director of Machine Learning Bhargav Avasarala, is to create a virtual open door policy. “Everyone is going to have meetings and be unavailable at certain points in the day, but the key to recreating those impromptu conversations that establish team chemistry is allowing remote employees the chance to chime in whenever,” says Bhargav. The Ever AI team uses Slack for most of its communication, but the team members are encouraged to dial up video calls whenever they want to connect with the SF team. The team also allows all other members to view their personal calendars to keep visibility universal across the company — and yes, that includes the CEO.

Be Liberal with Your Use of Public Communication Channels

Pierre Massat, Principal Engineer with Ever AI, realized that the biggest issues the team had early on were that employees at its headquarters forgot who was in the room for specific conversations. The failure to “see” remote employees resulted in muddled communication with remote team members who didn’t understand the context of follow up conversations via email or meetings. Ever AI discovered that it takes a commitment to conversation documentation to keep everyone on the same page. “99.9% of the time, it’s impossible to over-communicate — use Slack, use whatever you need to, but always put your work or thoughts in a public channel so that everyone can follow along.” Communicating in an open, public arena allows all team members to stay up-to-date. It also increases visibility for other teams within the company.

The importance of communication doesn’t begin and end with business talk. Tiffany also emphasized the relevance of bringing “water cooler” conversations into public channels. “You can’t replace walking to get coffee with someone at the office, but you can create opportunities for people to have fun and show some personality.” Using a random channel on Slack is a great way to facilitate this kind of banter. The Ever AI team frequently pops in hilarious family photos and weekend recaps.

Don’t Hire Leeroy Jenkins — Hire Engineers That Can Collaborate

When a company decides to go remote, they free themselves of the geographical talent challenges. But talent alone doesn’t make a good coworker whether that individual is sitting next to you or lives in another country. It takes a quality communicator to collaborate, showcase, and explain their work. For remote workers, this is crucial. Collaboration and communication are tied for the single greatest problem with remote work as reported by remote employees. The team at Ever AI factors this challenge into their hiring process by conducting the majority of their interviews online and asking questions that require the candidate to walk the team through their thought process. “We value people who can code and communicate equally well.”

Whiteboard in Person Or Invest in Online Tools

Ever AI might champion remote work, but they will be the first to tell you that some things are better done in person. That’s why the team prioritizes pairs or groups of engineers that work in the same location on specific projects. Currently, teammates in Vancouver, Toronto, and Kitchener-Waterloo have ownership over different projects. Pairing engineers together gives them the freedom to bounce ideas off each other during the ideation process in a way that is hard to replicate online, regardless of the investments the team has made in collaboration tools.

Impromptu whiteboard sessions are indispensable. “Coding is coding — you can do at any time in any place, but research and brainstorming are best done in the same room,” says Bhargav. The team concedes that pairing engineers together in multiple locations won’t work for all companies; after all, many startups have success with fully distributed teams, but it has been key to Ever AI’s success. Ever AI also hosts EverCon, a summit that brings the whole engineering team together as a full cohort 2–3 times a year to get big ideas and projects outlined.

Create Weekly Communication Outlets

Deep work is alive and well in the world of remote work. Remote teams benefit from uninterrupted time working alone. Those distracting side conversations, jokes, and outbursts don’t make it across fiber optic lines nearly as often as they do across the desk. Productivity is higher in this regard, but the lack of these interactions increases the importance of time spent together.

Most local teams have standups, share-outs, or other weekly meetings, all of which serve an even higher purpose with remote companies. Such activities inform, educate, and help employees interact. Ever AI has three such weekly meetings; machine learning lunches, weekly engineering meetings, and Lunch Bash (or snack hour for those in Toronto and Kitchener). The Lunch Bash is similar to how teams that focus on OKRs share progress updates and celebrate wins at the end of the week, but one of the main reasons Ever AI put it in place was to minimize essential questions that would go unanswered.

“We found that when we didn’t do a recap of what every team was working on at the end of the week, there were more questions — some of which wouldn’t get asked, and then people were left guessing,” maintained Pierre. “The Lunch Bash keeps everyone on the same page.” The meeting increases transparency across teams and gives the engineering team the opportunity to learn more about what their customers are saying and asking of marketing and executive teams. The Machine Learning lunch and weekly engineering meeting are meant to dive deeper into specific problems and projects and to give team members a chance to talk out what they have been working on.

To learn more about how Ever AI built its remote teams, visit: terminal.io/customers.