“How to Build Teams that Bring Together the Best of All Specialities”: Arathi Mani, Engineering Manager at Chan Zuckerberg Initiative (Video + Transcript)

March 18, 2023

Arathi Mani (Engineering Manager at Chan Zuckerberg Initiative) discusses the process of building successful multidisciplinary teams that bring together the best of all disciplines. She demos the cellxgene.cziscience.com/gene-expression tool to visualize uterus cell types in the spirit of International Women’s Day.


Like what you see here? Our mission-aligned Girl Geek X partners are hiring!

Angie Chang: We are here with Arathi Mani, who’s an engineering manager at Chan Zuckerberg Initiative. She works in the science division. I’m sure she’ll go tell you all about herself. I’m gonna just hand the mic over to her and say, welcome Arathi!

Arathi Mani: Well, thank you so much. I’m really excited to, to be here today. My name is Arathi. I am an engineering manager at the Chan Zuckerberg Initiative Foundation, and I’m excited to talk to you today a little bit about how we build multidisciplinary teams at CZI. Before I start, I just wanted to say please pop your questions, comments into the chat, into the Q&A. I’d love to kind of get that feedback. I won’t be able to quite see the questions with the wave my setup is, but I’m going to try to end a little bit early and then get to all of the questions at the end. Please, love an interactive session. I’m kind of sad that I can’t see all of you, so please go ahead and, you know, use the chat, use the Q&A.

Arathi Mani: All right. A little bit about myself and my team. I’ve been at CZI for a little over four years. I started out as a software engineer and then transitioned into engineering management, and I lead the single cell engineering team. I’ll try to make sure I get to all of the jargon, but at a very high level, single cell biology is the study and analysis of genomics, which is comes from DNA proteomics, which comes from proteins and transcriptomics, which comes from RNA at the individual cell level. And right now our focus is primarily on transcriptomics, which is the study of RNA.

Arathi Mani: Before coming to CZI, I didn’t know a lick of biology. My background is in computer science through and through. And so, you know, anytime I talk about science today, I really have to thank my, my team for being patient with me with teaching me so much about cellular biology. I just wanted to share a few photos of my team. We actually just came back from a week in San Francisco. Our team is actually primarily remote, half on the west coast, half on the east coast. And it comprised of many, many different disciplines. We have folks who have a PhD in molecular biology, people who have a background in physics, folks like me who come from a computer science background, and we come together to try and build and bring together science and technology in a way that really accelerates science forward. I’m going to talk a little bit about our mission and vision, just as a backdrop to have as a to have as a backdrop for all the conversations forward. CZI’s science mission is to cure, manage, and prevent all disease by the end of the century.

Arathi Mani: We are very ambitious. One of the things I love about CZI is we’re doing something a little bit unusual by pairing together science and technology. A lot of groups in the science world are primarily kind of academic labs, or they might be part of the biotech industry which, which tend to be for profit. And the cool thing about where I work about the CZI Foundation is that we are a non-profit foundation. All of the software that we build is really focused on trying to accelerate the cutting edge work that a lot of the academic labs are doing, where technology can really help make that go faster, make their workflows go faster, and we build everything as open source software. We don’t sell it or anything like that. Iit makes for a really unique organization with a diverse set of folks who are trying to all kind of push this vision forward.

Arathi Mani: The single cell mission is to create knowledge about what are the mechanisms at the individual cell level that cause human disease. And then by making that knowledge accessible and available for all scientists to then accelerate the generative generation of curative treatments. Our vision is to create a reference of human biology. Our goal is to try and sequence all of the different cell types in the human body and make that data accessible for anyone to use. And then also pair it with visualization tools to enable, you know, biologists and physicians who may not know how to work with big data to be able to quickly get insights out of different cell types and to understand the cause of each disease. We have a very kind of big mission and vision, very much rooted in the science.

Arathi Mani: We do have a multidisciplinary team to try and push this forward. I’m actually going to just jump to the punchline of this entire presentation because I want you to have this this notion in your head as you’re listening to me speak. But I think the key thing about what makes for a successful multidisciplinary team is all about empathy. It’s this idea that if you try to understand where other people are coming from, to try to understand their culture, to understand their background a little bit and walk a mile in their shoes if every person on your team does that, it’s what creates cohesion. It’s what creates a better environment from which you can have discussions and make decisions and really push push something forward and create something that is, you know, bigger than the sum of its parts.

Arathi Mani: I hope you keep this in mind as I continue to talk throughout this presentation. I wanna talk a little bit about the challenges first. I know that not everyone here is working in a very science focused domain. But I do think probably all of you all have been working with at least one individual that comes from a different background than you. It could be as simple as just engineering and product management or engineering and design or, you know, any kind of combination. I hope you kind of get a little bit out of this presentation, even if the scope of the multidisciplinary work isn’t as broad as it is at CZI in order to achieve our mission for every single project we may have individual that come from upwards of five specialties.

Arathi Mani: They could come from engineering, product management, user experience research, product design, computational biology product analytics, so many different specialties that are all involved into making one project successful I kind of see three different challenges that can manifest. One is that each specialty can come from can have a different sense of what is important to them. What are their culture and values? And they may not recognize what that other specialties culture or values might be. There might be just like general lack of awareness about what the different, what the differences are.

Arathi Mani: The second thing is around language barriers that can result in miscommunication. As somebody coming from tech, I think I often will accidentally use a lot of tech jargon forgetting that not everyone in the room might necessarily know what I’m talking about. I definitely felt like that when I first joined CZI. And, you know, a word like transcriptomics, I would sit there and be like, I don’t know what that means. You know, it happens very often where you sit in a room and somebody’s speaking and you don’t quite understand exactly what they’ve said because we’re so used to using our own jargon.

Arathi Mani: And then the last little bit is that the responsibilities at the intersection of the specialties can be a little bit murky. When specialties need to collaborate very, very closely together, it can be sometimes a little bit difficult to understand who does what. And I’m gonna spend just a couple of minutes going through a specific example, kind of demonstrating two ways that building this particular tool where we had some of those challenges. I’m gonna switch over to sharing another tab, which I hope you all can see.

cellxgene cziscience gene expression tool
cellxgene cziscience gene expression tool

Arathi Mani: This is a tool that enables folks to to understand what are the cell types in any tissue, and what are the genes that make that cell type unique?

Arathi Mani: In the spirit of International Women’s Day, let’s pick a very cool tissue – the uterus! You can see here all of the different cell types in the uterus. Uterus is also an incredibly cool muscle. That’s what enables it to expand from the size of a lemon to the size of a watermelon. And I’m gonna add all of the marker genes to the plot.

Arathi Mani: Marker genes are genes that are specifically unique to that particular cell type, which in this case I picked a muscle cell. You can see a plot rendered here, and if you’re a scientist, you can start to understand, ah, these genes are specifically unique.

Screenshot at .. PM
Visualizing uterus cell types at cellxgene.cziscience.com/gene-expression tool

Arathi Mani: There are certain genes that, that may not be unique or are prevalent in other cell types. Some of the challenges in getting this tool out the door. One thing was figuring out the algorithm to actually compute this.

Arathi Mani: We had a computational biologist who went out into the field, did some literature research to understand what the algorithm should be. Ultimately it was a software engineer who implemented it. There was a kind of a discussion about like, who owns this algorithm moving forward? Who owns the the writing of it, but then the long-term maintenance of it like how does that exactly work? And even trying to figure out who writes the first draft of the algorithm was a point of discussion.

Arathi Mani: Another kind of discussion that happened is, when was this good enough to ship? When was it good enough to, you know, remove the feature flag and make it available for all scientists to use? We had very different perspectives from our computational biology folks saying, oh, we need to kind of validate this against scientific literature to make sure we’re not accidentally you know, making bad science available.

Arathi Mani: We had product managers who are kind of in the middle saying, we should validate it for, you know, certain popular cell types, but don’t need to do it kind of comprehensively for every single one. And then you had engineers who were like, well, the unit test passed the smoke test pass. We’re good to go and let’s ship it. We had a very, very different you know, theories on when this was good enough to ship. All right.

Arathi Mani: Going back to the presentation – so how do we get over this? I think, you know, number one thing is culture building. There are different aspects of culture building. One of the things I really love about CZI is the clarity in our unifying mission and vision. I think having that vision and mission is what sets the groundwork to have productive disagreements, because it kind of gives the team confidence that everyone in the room has the why behind their work.

Arathi Mani: As a leader, if are in the position of being a leader, but even if you are if you’re not, you want to be able to repeat and re-articulate this mission as as often as possible and make sure that everyone really understands it. If you have the same why, that’s your common framework, and you have a, you know, a baseline.

Arathi Mani: Establishing a strategy is the kind of second thing. A strategy should be clear, easy to understand. I think at the start of developing that strategy, you should make sure you have generative conversations where people have a voice and have an input into the strategy. And I think it should happen across all different levels. And then ultimately, when that decision is made on what the strategy is, it’s okay to say like, not everyone not everyone may agree with the approach, but it is important that people disagree and commit. And it kind of goes back to setting that groundwork so that if people understand the why and what our approach is going to be, then you have at least a baseline from which you all can start discussion.

Arathi Mani: And then the last thing is around team values. Establishing and intentionally establishing team values is incredibly important. It helps keep everyone on the same page in terms of how we wanna work and what do we care about the things we build. The conversation about what is good enough to ship was a great conversation to have. And out of that came kind of a set of values from which other projects can use to say, ah, like, this team did it, you know, establish this particular values for this set of values for what is good enough to ship. We can kind of borrow that and move that forward for future projects.

Arathi Mani: The next thing is around expectation management. It is really important to spend the time to kind of define roles and responsibilities. And then in this kind of remote working environment, documenting them is, is really important to kind of alleviate any issues around miscommunication. And then even once you have that at a high level, establishing who does what ahead of a project I is really important. The conversations around who develops the algorithm and who does the maintenance of it was very productive to have, even if we had it a little bit later in that project, once we actually stumbled across that that, that blocker, once you have that and documenting it, people are on the same page that can really help in, in fostering a positive team culture. And then if you’re not sure who should do it, I really think that people should feel empowered to try it themselves and really feel ownership over the, over the project that they’re do, they’re they’re working on.

Arathi Mani: And that kind of goes to the punchline of, once you’ve actually established all of these roles and responsibilities and who’s doing what, then find opportunities to cross the lines and break the rules. You know, be thoughtful about how the way, the way that you do it you know, you should have the conversations with, with folks from other disciplines. But it’s a great way to build empathy and to understand the culture and the perspective of somebody else by kind of stepping over the line a little bit. And for example, in this case, having a software engineer develop an algorithm or implement an algorithm that is from scientific literature. And then the last thing I wanna talk about is communication. And this is kind of, I’m gonna talk about things that are a little bit very much tactical here. One thing that we did at CZI recently is delete all Slack channels that were created for specific specialties and only have Slack channels that are for whole projects or whole teams.

Arathi Mani: The reason that we did that is that I really believe that you can create and really foster cross pollination and the sharing of ideas. We had a Slack channel previously just for computational biology where they were sharing papers, but now it goes into a broader team channel where, you know, folk engineers and product managers and designers all have the opportunity to see what papers, what scientific papers a computational biologist thought were interesting, and make an attempt to read it. And then, you know, in my role as an engineering manager, as I develop goals for my reports, creating incentive structures for individuals to learn about other domains is really important. Making sure that it becomes part of their goals, making sure that they have the time to go and try and read a a paper or go and attend scientific conference.

Arathi Mani: I think making space for that is incredibly important. And again, kind of go, goes back to building that sense of empathy. I’m going to try and quickly wrap up here, but I hope you heard this throughout the presentation, this kind of focus on empathy and trying to build that in many different ways as much as possible throughout your entire team. I think that this really is the key aspect of building a successful multidisciplinary team. With that, thank you so much for, for listening. And I’m gonna switch over to try to see what’s in the q and a and chat and answer some questions. I love, I love the comments here. Very, very cool background in clinical data management and program management. That’s, that’s awesome. Well, thank you all so much. It’s, thank you so much for the the interaction. I’m so sorry that I you know, it’s not to face, it’d be so cool, but yeah. Thank you.

Angie Chang: Thank you, Arathi. That was an excellent talk. All right, we’re gonna break now and go to employer booths! We have Autodesk, Cadence, Dematic, United States Digital Service, and CodeSee in employer booths. If you go back to look at the schedule, you can click on the link to go to a booth, or you can go to the navigation to the top and click on employer booths and just find them there. They’ll be there for the next hour till one PM Pacific. And yeah, thank you so much again. All right, see you at one for the keynote. All right, bye.

Like what you see here? Our mission-aligned Girl Geek X partners are hiring!

Share this