Q+A: Building a data science and engineering team for the social sector

Several weeks ago, IDinsight was named one of 15 orgs receiving support from Google.org as part of its company-wide commitment to achieve the SDGs with AI solutions. This project is just one of the partnerships managed by our Data Science Engineering and Monitoring (DSEM) team. We sat down with Fola Adegbemle, Manager – Research Evaluation and Data team operation to hear about how he has helped grow a social sector team with private sector expertise.

Q. What are you most proud of since joining IDinsight?

While I think we’ve done pretty well regarding achieving most of our strategic goals, I have found the manner we’ve gone about it most impressive. I am proud that we have not compromised on any ideals on that journey.

We’ve managed to (almost) triple the size team in a fiercely competitive market for technical talent and stuck to our organizational values along the way. Our approach was to 1) Hire team members who are mission-aligned 2) not compromise on standards of technical excellence, 3) hire team members who are representative of the contexts in which we work, and 4) ensure that the team is diverse.

To paint a picture, we now have 24 team members from nine countries, mostly based out of our hubs in Nairobi and Delhi. Greater than 70% are regional nationals of places where we work, and 30% of the team is female – while still falling short, it is above industry averages for similar technical teams.

Q. What do you think has contributed to achieving these goals? 

Many of our social sector partners do not have the in-house expertise or resources to apply big data to solve their most pressing challenges. Those abilities remain exclusive to the private sector in many countries, and I see the DSEM team’s work directly filling that gap.

In hiring, I think it helps that our work connects to a larger purpose: enabling broader access to world-class advanced analytics that would otherwise be inaccessible in the social sector.

Q. What approach did you take to grow the team? 

We realized, looking at our project pipeline, that the demand for the DSEM team’s services would outstrip the current team’s capacity and we would need to expand. In the beginning, the team was comprised of three senior leaders who led the major verticals within the team and a couple of junior engineers. We immediately prioritized building a managerial or middle layer alongside our recruitment drive for more individual contributors.

We want to be methods agnostic so we can pick the method/tool that is the best fit for a partner’s problem. So, we focused on hiring engineers with a solid foundation in applied math, statistics, and computer science. This allowed them to be flexible and pick up new methods easily. 

Over time, as we have learnt more about the demand for various services, we have hired specialists to grow our capacity in some focused areas.

Q. Given the radical focus on impact at IDinsight, how did we decide how much social impact experience was necessary for each role?

Unlike other roles at IDinsight, we certainly did not have the benefit of being able to recruit from a pool of candidates with significant sectoral experience, as data science and engineering in the social sector is still relatively nascent.

Rather than using a candidate’s previous experience as a proxy for social impact orientation, this was something that we explicitly sought to test as part of our recruitment process through a combination of screening questions on the application and our interview. So, while everyone did not have previous social impact experience, they were all extremely motivated to have a positive impact. In a traditional data science position, your work may be to increase ad sales by 1% or click-through rate by 2%. Doing something meaningful and interesting is a big reason by people come to us.

In fact, a lot of our engineers found us. They reached out because they saw the work we were doing and were excited about using their tech skills for good. 

Q. In hindsight, what do you think worked well?

I don’t consider myself a recruiter or a talent acquisition specialist, at least in the traditional sense. Still, there were several tactics that we used that were effective: 

• Focusing on our vision for the team structure: when I look back to many of the conversations I had with prospective team members, many found the vision that we had in place for a mature advanced analytics team very appealing. Joining IDinsight would not mean leaving technology to come and be the sole engineer or scientist at a non-profit – which they saw as a prospect that could stifle their growth despite the remarkable upside in impact potential. I credit much of the thinking around the vision and structure to Ben Brockman, who founded the team – and current leadership – Marc Shotland, Sid Ravinutala, and Eric Dodge.

• Understanding who we are competing with for talent. Having lived and worked in start-ups in Nairobi (which we chose for our second hub) for many years before IDinsight helped me immensely. For one, I knew many talented people working in tech, but more importantly, I understood the market. That helped us calibrate what we were looking for with what was available.For example, It was really instructive for the team’s leadership to know that most of the experienced talent that would come through our pipeline would be mostly self-taught, as most universities in the region are just catching up to the demand for the type of skills needed by our team. We were more flexible in evaluating things like impact orientation. And sometimes, that looked very different based on people’s backgrounds.

• Identifying what worked for us to bring in talent. Early on, we tried a variety of strategies to build a robust talent pipeline. Many seemingly great ideas, for example, even top search firms that had successfully found talent for other roles within IDinsight, did not yield any promise. It was great to find that out early and pivot our recruitment approach.

• Establishing beneficial relationships where we have found talented colleagues. We’ve worked with many amazing partners across the continent, such as Zindi, African Institute of Mathematical Sciences, Moringa School, and The ROOM, to understand the African talent landscape. For example, some of our earliest candidates were partner referrals who gave us early insights into candidate archetypes that would later make up most of our organic pipeline. Those valuable insights made a difference. 

• Crafting a delightful recruitment experience; In my past positions at ALX and, subsequently, ALU, there was a big focus on crafting delightful experiences for applicants. At ALU, under the leadership of the very brilliant Justin McDonald, we even had a running “design experience” workshop series known as Delight by Design instituted to keep design principles top of mind. And it made sense – if our process delighted our applicants sufficiently, they would want to join us.
  
  We operate with a similar mindset at IDinsight and enjoy reading positive notes from applicants who do not even make it through our recruitment process.
  

• Always being open to talent and having zero waste: Here is a secret: we run a rolling recruitment process for most recurring roles on our team. We don’t believe in hard deadlines and stage-by-stage cohort recruitment, We are always open to the best talent regardless of when they enter our pipeline. You know how everywhere you apply for a role unsuccessfully, they tell you they’ll keep your CV on file and alert you if any opportunities that could fit your profile arise? We actually do that! We think about how every role relates to a previous one that we hired for so, every search often begins with an analysis of the pipeline for previous roles. Previous applicants are such an underrated treasure.

Q. Post-recruitment, what challenges have you faced as a team?

Having an embedded, fully-fledged Data Science and Engineering team at IDinsight has meant, to some extent, rethinking how we collaborate internally (with client-facing colleagues) and how we work with clients. For example, our traditional project staffing models have evolved to incorporate new DSEM roles, which our client-facing colleagues often grapple with. We have very different ways of working from the rest of the organization. As techies, we work within an agile framework, which involves feedback and re-prioritizing.

Also, I think, and maybe more importantly, the expertise that the DSEM team brings to IDinsight is not something that many colleagues at IDinsight previously understood and vice versa. Hence, there’s been a need to constantly educate (both ways) on how our work complements each other in the service of our mission to use data and evidence to improve lives.

Q. Finally, what’s next for you and the DSEM team?

The reward for all our hard work is that we now have a fantastic team brimming with energy and purpose, and we are constantly on the hunt for the most impactful opportunities that could benefit from our capabilities.