Leading a team in a domain you are not familiar with

Problem

A few years ago, I inherited a small Machine Learning (ML) team. However, my existing teams were either in infra or in products and I was not familiar with ML. No one in my existing teams knew about ML either. I had so many questions. How could I know what was going on in the team? How could I support each engineer? How could I provide direction? These problems baffled me. To make it worse, I needed to scale this team up immediately.

Actions taken

I realized that all my challenges boiled down to one single problem. How could I establish my own credibility? This is the most important thing, as credibility fuels respect, trust, confidence, and energy. If I was credible then I would be able to lead the team well. My own manager also didn't know my domain. However, he led his teams very well, because he had credibility. I also realized that credibility could be earned in many different ways and that technical competence, despite being important, was only one way. I didn't try to ignore the importance of being technically competent, but I wanted to leverage the other ways to buy myself more time. First, I enrolled in a Coursera ML specialization. This was a large commitment, which involved taking a bunch of courses. I set aside one hour per day and acknowledged that it would still take me at least a year to finish the course. But I needed to establish my credibility sooner. So here is the second thing I did. I asked the experts on this new ML team to teach me. I asked them about the basics while discussing their project and I also asked them to point out the resources that would be useful for me. By asking the ML team to be my "mentors", I increased their engagement, while also increasing my credibility. The third thing I did was to lower my status and be honest with them. If I didn't know something, I acknowledged this. If I made a misjudgment, I acknowledged it immediately. This greatly improved the trust in our relationship and made them feel that I wanted to help rather than to rule. The fourth thing I did was to instigate a few large, cross-team projects. This stirred the team's internal dynamics and made people on the team more open-minded. It also provided the team with a kind of "brain food", that it otherwise would have needed from me. Moreover, the feedback from other teams also provided good reference points on how well the team was doing. All the above steps not only enabled me to focus more on recruiting but also enabled me to identify two high-potentials from the team whom I could leverage to scale up the team. Within only six months, this team became one of the highest performing ML teams in the company, and it doubled its size too. I, myself, also finished the ML specialization within less than a year.

Lessons learned

As a manager, your effectiveness depends on your balance between warmth and strength. First of all, you need to be a caring person with character. This is warmth. Sometimes, however, your strength may be threatened. For example, when you inherit a team in a domain that you are not familiar with, your balance will become lighter on the strength side. In this case, you need to increase your strength. That's where credibility comes into play. There are many different ways to increase your credibility. Don't limit yourself to only technical competence. Although technical competence is important, you will need other ways to buy yourself more time. That's where your warmth can help you. Sounds interesting, right? You can use your warmth to help you increase your strength. Be authentic. Be true. Be curious. This will help your team to see your leadership strength, while you catch up with your technical strength.