Forward-Deployed Teams — A Catalyst for Public Sector Digital Transformation
A short introduction to DSAID’s agency-facing data science teams.
Organisations in both the private and public sectors recognise the importance of building a data-driven culture and the need to incorporate data science into their decision-making and operational processes. Yet for many, getting there is often not trivial. Answering these challenging questions is a key part of what I do as a GovTech forward-deployed data scientist.
What exactly is a “forward-deployed team” or “FDT”? Well, you can think of it as an outpost. Although we are proud members of the central Quantitative Strategy (QS) team under the Data Science and AI Division (DSAID), we are embedded directly within other Singapore government agencies. Our role is to accelerate analytics adoption at the host agency. In my own experience at Enterprise Singapore (ESG), we do so by undertaking strategic analytics projects and promoting best practices in data science in order to help the organisation transform digitally.
Our FDTs typically comprise two to three data scientists. In my case, Jing Song and Shane are my comrades at ESG. But how does such a lean team shore up the analytics capabilities of an entire organisation? Can it really scale?
Plug and play where possible; call a friend if needed
Although the FDTs are the face of DSAID at the agencies, behind the scenes, there is a larger brain supporting us in the work we do.
Two of DSAID’s greatest assets are the interesting pool of reusable resources that the team has put together over the years and the diverse skillsets of our people.
Need a script for doing optimised reverse geocoding?
Amit has a Git repo for that.
Looking for a dashboard that perfectly displays to users the performance of your machine learning models over time?
Grab it from Rory.
Venturing into model deployment?
Check in with the DSAID product teams to find skillsets in DevOps and software engineering.
Need to find out what data from Ministry of Manpower can be used to support your analysis?
Talk to Lionel!
The ability to plug and play these shared resources allows the FDTs to save time on things that DSAID has already developed as a team and to focus on customisation and innovation at the agencies.
But beyond reusable repos, what has been a strong bedrock for me is the extensive network of technical and domain expertise in DSAID that I can tap on at a moment’s notice. All this is only possible when we have a supportive and collaborative culture that seeks to nurture its people.
We equip and journey together
“If everyone is moving forward together, then success takes care of itself”
— Henry Ford —
Analytics adoption should not begin and end with just an analytics team; participation and a shift in mindset must happen at every level of an organisation for change to be sustainable. At ESG, we collaborate closely with the agency’s planning division to co-conceive a hub-and-spoke data champion structure. In short, analytics enthusiasts from each ESG division are trained and empowered to initiate and deliver their own analytics projects while being guided by the central analytics team, the FDT.
To-date, we have trained more than 100 data champions, many of whom have gone on to initiate and build impressive dashboards and models that are currently supporting operations within or across divisions. Having started as a team of two data scientists, it is heartening to see many new like-minded and competent companions who are now walking alongside us.
What does being part of a FDT mean to me personally?
Plant a seed and watch it grow
The ESG outpost has evolved considerably over time. Projects wise, our FDT predecessors trail-blazed by delivering many useful data analytics proof-of-concepts to drum up interest in the use of data and by laying the groundwork for our data champions structure. Carrying their torch and building on their foundation, we are now moving towards more complex solutions that can be deployed into applications to support users’ decision-making.
This process is not without challenges. As ESG’s analytics maturity and appetite have expanded over time, Jing Song, Shane and I have had to pick up new skills along the way, including ones that go beyond data science. Examples of these are data engineering, DevOps and Cloud computing skillsets, as we deploy use cases into production with the strong support of our GovTech ESG IT team. Nevertheless, we thoroughly appreciate the opportunity for personal development and to grow together with ESG. What’s more amazing is that as we provide fertile grounds for analytics to become more accessible — that’s when individuals and departments in ESG start to develop technically in their own ways.
In terms of how we scale up the agency’s data science capability, we are moving from doing ad-hoc projects to designing an ecosystem for analytics to thrive. This includes driving the creation of more specialised roles to take charge of essential parts of the end-to-end data analytics journey such as data governance, data management, data democratisation, ML engineering etc. As the FDT’s capability development work takes root in ESG, I am excited to see how our work will contribute to ESG’s key economic decisions.
There is always something new to look forward to
As someone who enjoys trying different things, the varied experience that comes with the FDT role keeps me excited about my current and future work.
Deepening domain-specific knowledge. FDT’s projects are very targeted at the agencies’ operating context. This allows us to grow not just technically, but also in our appreciation of the host agencies’ business objectives. Having worked closely with our ESG stakeholders, I have a front-row seat in witnessing the development of our local enterprises. I feel great joy seeing the analytics products designed by the ESG FDT and data champions coming to fruition, supporting the agency’s mission. To quote a recent example, the team is in the midst of implementing a text mining project that we hope will help ESG track the mounting enquiries that our local businesses are reaching out to ESG for during the Covid-19 period.
Reaching for breadth. All FDT data scientists would return ‘home’ to the central QS team after their 2–3 years stint. This means that we can grow deep in a specific domain during our deployment and broaden our experience as we take on cross-agency, consultancy-type of data science projects at the central team. To me, it is a unique opportunity to experience both an in-house and a consultant-like role all in the same job. And more than that, many past QS officers have also rotated to other teams in DSAID to try new things, new roles and to broaden their skillsets. So, if you are also looking to find your heart’s calling by growing deep domain expertise and yet retain the flexibility to explore a wide range of different projects, explore options to be deployed to a FDT.
I hope you now have a better idea of what FDT data scientists do. Ready to make an impact in the Public Service through data science? Check out Lin Zi’s POST on how you can become a QS data scientist!
