Building Technology to Map Every Cell in the Human Body: Arathi Mani’s Engineering Journey

Charles de Bourcy
CZI Technology
Published in
6 min readJan 6, 2020

The only way technology can solve the world’s most pressing problems, like improving education, managing disease and reforming the criminal justice system, is through radical collaboration between technologists and those working on the frontlines of these very issues. For many software engineers, building products toward our missions in Science, Education, and Justice & Opportunity is quite different from what they’ve worked on before. We sat down with software engineer Arathi Mani from the science team to learn more about her journey to the Chan Zuckerberg Initiative (CZI).

At NeurIPS 2019

Arathi, what was a pivotal moment for you in your career?

As far as I can remember, I’ve generally known what I was interested in and what my “next step” was. CZI was a big step into the unknown for me, so that was actually a pivotal moment!

Prior to joining CZI, I was happily working at a large, well-known tech company, though I was looking to get closer to my interests in Natural Language Processing by switching teams. Coincidentally, I received an email from a CZI recruiter and something made me stop for a second instead of dismissing it right away. I first heard about CZI through Mark and Priscilla’s kickoff Facebook post in 2015 and remembered that their mission was quite ambitious and intriguing, so I decided to take a second look.

Taking that leap of faith between something I knew I could excel at and something I really didn’t know much about was hard! I actually bought a picture book on biology the day I signed my offer letter. At the end of many days of consideration, I came to CZI for two reasons: mission and passion.

CZI’s mission in science is incredibly ambitious: to support the science and technology that will make it possible to cure, prevent, or manage all diseases by the end of the century. What’s even more incredible is those who work here believe this mission is achievable and are earnestly working toward it.

What surprised you the most when you came to CZI?

Many places express the notion that they have a collaborative working culture, but I was struck by the essential place collaboration occupies at CZI. I work on the Human Cell Atlas (HCA), which is a worldwide effort to create reference maps of all the cells in the human body as a resource to better understand health and disease. On this project, I work with folks who are in two different countries, three different time zones, and four different institutes. In no other project had I ever experienced the amount of coordination and cooperation required to organize a team this diverse.

In addition to collaboration within an engineering project, collaboration across domains is fundamental to the way we work at CZI. Technology is wonderful and can bring about amazing change, but including domain expertise when discussing solutions is imperative. We have an entire team of computational biologists providing so much insight and interpretability to the tools that software engineers build. I love the fact that we see the importance of such collaborations and have the opportunity to set examples for future work in multidisciplinary fields.

At the 2019 Grace Hopper Celebration

How does the need for scientific collaboration translate into the technical challenges you’re working on?

I recently worked on a collaborative project involving metadata about the datasets that scientists are contributing to the HCA. Metadata is a notoriously tricky topic, as different scientists have different standards for what metadata should be collected and how the collected metadata should be organized. In the HCA project, metadata is broken up into modules (and embedded sub-modules), and each module/sub-module has an organizational standard that is designated by a schema. In total, the HCA has over 70 metadata schemas with logic about the schemas embedded in our data coordination platform software system.

Technologies that can characterize individual biological cells are rapidly iterating and improving, so the subsequent metadata that arises out of these experiments changes swiftly. This raises a tricky technical challenge, as our metadata schemas need to be flexible to absorb rapid transformations and then ensure that they propagate throughout the rest of the system, particularly in downstream modules that serve data (like cell-by-gene matrices) to data consumers. In order to allow for rapid metadata schema changes, I built software to allow downstream system components to pin schemas, enabling the decoupling of metadata schemas from other modules of the software system. I also wrote software to generate automatic tests based on the changes that occurred in the schemas and notify consumer-facing modules if their subsystem was broken.

This project required the coordination of all teams working on the HCA infrastructure to ensure that every component that had dependencies on metadata schemas had either built a mechanism to handle and retry errors that arose from changing metadata schemas or had chosen to make use of the pinning system. We had many discussions leading up to the implementation. It was so important to make sure everyone was on board with the new set of Standard Operating Procedures (SOPs).

With members of the CZI Science and Computational Biology teams at NormJam

Which of your past experiences has best prepared you for working here?

Humility and openness is vital when working on multidisciplinary teams in order to facilitate convergence to a solution. In traditional tech companies, being extremely proficient in one or two particular verticals of computer science will often serve you well; at CZI, engineers are partnering with folks in policy or education or science, so solidity in a particular area of expertise is not always sufficient. We need to be humble enough to listen and learn from domain experts to truly understand where technology is actually useful to further our initiative areas.

To that end, I think that my teaching experiences (as a camp instructor for children as well as a lecturer for a college-level class) have served me well to prepare me for my role. As a teacher, I could prepare as much as I wanted for a lecture or a class, but at the end of the day, I needed to listen to my students to understand where the gaps were and focus on that instead of regurgitating a lesson that didn’t address the individual class’s needs. I learned to hone personal traits such as patience (nothing like 7-year-old campers on summer vacation to teach you that skill!) and learned to improve my communication through these experiences. These “soft” skills have truly allowed me to thrive at CZI.

What is something people wouldn’t know about you?

I love making cakes! I volunteer with a charity called Cake4Kids, which partners with various shelters throughout the Bay Area to bake birthday cakes for children who might not otherwise get one. It’s a very fulfilling way for me to give back to my community while getting to do something I love.

We’re hiring! Is your dream job here? Check out our openings.

--

--