In this edition of Getting to know our DataLeaders Rohan Samaraweera shares his career journey with data and the experience that has led to his current role as a Senior Director of the Data Science Branch at the Department of Home Affairs. Rohan highlights the way his team is leading data science efforts at a time of significant technological change within Home Affairs and more widely across the APS.
My career journey into data was founded on a simple twist of fate – a chance meeting with an Australian Signals Directorate (ASD) recruiter at a university careers fair. Throughout my time at ASD, I gained a great deal of experience and learned many lessons that continue to shape my thinking and approaches.
ASD was a fascinating place to work as it straddled the nexus between data, technology, international relations and security. In-house tradecraft was unique, requiring blended approaches drawing on its workforce's diverse talents, strengths, expertise, and life experiences. Close collaboration between linguists, intelligence analysts, engineers, developers, ADF personnel, mathematicians, signals analysts was a daily routine. What linked us all was the 5Vs of data – Volume, Value, Velocity, Veracity, and Variety.
Initially, value, volume, and velocity were our primary challenges. ASD needed to use limited technical and human resources to quickly find the most valuable foreign secrets without crashing ICT systems. My role was to curate haystacks and find the needles hidden within. I learned to use business intelligence software to link disparate datasets together, provide statistics on performance, and make a case for resource decisions. At that time ASD was on top of its game, often one or two steps ahead of the day's technologies. However, disruptive change was soon on the horizon.
The rapid advancement of digital technologies was both a blessing and a curse. On the one hand, we benefited from immediate improvements in computing speed, falling costs of data storage, and the explosion of available data. On the other hand, new technologies also levelled the playing field and changed the very nature of human communication. This issue, coupled with dramatic shifts in our strategic priorities necessitated dramatic change.
Combating global terrorist networks, enabled by increasingly cheap and ubiquitous digital communications technologies, required a surge in the spread of data analysis skills, constant upskilling and innovation. Data variety and veracity grew in importance while cyber and policy specialists became increasingly critical collaborators. Data analysis and storytelling played an increasing role in intelligence production. We used network graph analytics and geospatial tools to make sense of target networks and derive insight into activities.
ASD owed much to its wartime heritage: an integrated civilian/military workforce; commitment to a single clearly defined and succinct mission; and, of most resonance to me, the idea of growing in-house talent.
Early in my career, an ex-military colleague gave us a presentation on the importance of investing in staff and creating a learning culture, which I still practice today. He explained that:
- an untrained employee negatively affects productivity;
- in physical vocations such as the trades, a master artisan can be four times as productive as a newly qualified apprentice;
- in intellectual professions that gap can be tenfold; and
- in intellectual professions, when technology is added, that gap could be as much as 100 fold.
These days with artificial intelligence (AI), that gap could be 1000 fold or more – limits determined by the task's magnitude, not the resources put forth to perform it.
I became a devotee of this thinking, volunteering as adjunct faculty for ASD's training academy, developing training in data analysis and communications technologies.
I now lead data science efforts within the Department of Home Affairs at a time significant technological change. Modern government institutions built upon the written word increasingly need to accommodate numbers and data to maintain relevancy and public trust. Home Affairs is a big, diverse agency with a broad range of responsibilities, many different functions, areas of specialisation, and a vast array of disparate data assets.
Driven by challenges and opportunities identified through collaboration with business areas, we work to incrementally expand our repertoire of tools and techniques into new areas such as computer vision and natural language processing. This effort is backed by investment in our people, including through a pilot micro-credentialing program. We work closely with our IT colleagues to improve access to data, technology and infrastructure, and explore new opportunities such as cloud computing. Our work is increasingly enriched, refined and amplified through collaborative partnerships with domestic and international partners in the public, private and academic sectors.
Building trust in our capabilities is a top priority. We have invested significant effort in lifting awareness of data ethics, and improving transparency and accountability, learning lessons from other agencies. ASD built compliance into their culture by ensuring all staff understood the legislative foundations underpinning their work and were trained to apply sound judgement when navigating the inevitable 'grey' areas. Home Affairs has commenced a similar journey by developing data ethics and algorithmic bias training for our staff and stakeholders. It is the first step towards ‘baking in’ ethics and integrity into our culture and practices.
Government institutions should be a key player in the age of AI as they have several critical ingredients required to excel. They are the custodians and collectors of vast amounts of unique data and house deep and diverse human knowledge and expertise across many domains. Innovation has historically occurred through intersections between people working in different domains coming together to solve problems or explore new opportunities. In the APS there are well established inter-agency relationships, enabling collaboration, partnership and pooling of resources to create public good.
It is an exciting time to be working in data, analytics and AI in the APS. Data and technology will increasingly augment the APS workforce and, with sound execution, create significant public good. Right now, the APS is laying critical foundations to ensure a positive transformation pathway and initiatives such as the APS Data Professional Stream are helping to facilitate progress. I encourage you to develop your data skills, collaborate, network and learn with others, not just within your agency. Technology is undoubtedly a significant enabler but ultimately it is people that are critical to Australia's data-enabled future.
Rohan Samaraweera
Senior Director, Data Science Branch, Department of Home Affairs