Skip to main content

Minimum Viable Services for UK Taxpayers

Kalbir Sohi (drie)
Average rating: ****.
(4.40, 5 ratings)

HMRC is the UK’s tax authority and offers online services to millions of UK taxpayers. Over the last 12 months we’ve moved to delivering services by putting out alpha and beta versions of new products and testing hypotheses about them. This has been very successful from the perspective of developing and improving user facing features but in the infrastructure and operations space we’ve found it harder to convince people to put out minimum viable services and test hypotheses about them. In this talk Kalbir will look at some specific cases where we have had success and failure in adopting a hypothesis testing approach to infrastructure and operations design including how we’ve:

  • Done 24/7 support
  • Managed security needs
  • Scaled to support 20 delivery teams
  • Use of new technologies such as docker and heka
  • Worked with the demands to C*O level stakeholders

For each of these areas Kalbir will give an example of where we’ve tried an approach of delivering a minimal viable service and iterating on it; and whether that has succeeded or failed (maybe for the right reasons!).

Kalbir Sohi


Did a PhD in philosophy, then started to do technology in government. Been working as an infrastructure product manager in HMRC for 8 months, led a team of 10 engineers to build HMRC’s first web application platform on cloud infrastructure in that time. Aiming to grow it to serve millions of UK taxpayers.