Engineering in Asia: lessons from scaling Stripe engineering in Singapore

For an overview of the region, check out my previous post: Observations from APAC.

I recently relocated to Singapore to kickstart Stripe’s new engineering hub, but prior to that spent almost ten years scaling engineering and product teams at both Facebook and Stripe. I came into the role excited to apply my previous experience to an entirely new market. While a lot of what I know still applies, the differences have been even more surprising than I expected.

Local engineering salaries can come in at a fraction (¼ — ⅕) of US averages, while teams can quickly expand from 0 to thousands in just a few years. The big giants (India, China) have orders of magnitudes more engineers than the rest of Asia combined and operate very differently as a result. Experienced engineering management is a precious commodity in Southeast Asia.

Note: While I’ve met engineers and founders from all over the region, most of my time has been spent with people in India and Southeast Asia. I have yet to experience tech culture first-hand in places like China and Japan, but am very excited to visit soon.

Engineering talent is universally valued; engineering culture is uniquely different

People see the value of developers and the part they can play in the economy, but engineering culture varies widely depending on the subregion. China and India are home to hundreds of thousands of developers, have established practices, and are seen as talent pools for Southeast Asia. Given the density of talent, they also tend towards localized development centers with heavy reliance on in-person communication, sometimes complete with rows of cubicles as I saw once on a visit to Bangalore.

On the flip-side, Southeast Asian companies go distributed quickly. Most of the companies I’ve talked to have engineering teams in multiple countries. Locations are based around acquisitions, seed leadership hires they can build a team around, or just where they can hire faster.

  • GO-JEK has made 9 acquisitions to date (including 3 payments companies) and the engineers from these acquisitions now total ~50% (!) of the overall team.
  • Grab was started in Malaysia, quickly expanded to Singapore where it’s one of the largest engineering employers today, and is actively hiring in China, India, Indonesia, Vietnam, and even Bellevue, WA.
  • Stripe’s newest engineering hub is in Singapore, we’ve hired people from multiple countries in the region already, and will support exceptional remote engineers in other APAC offices as we grow. (Stripe has supported remote engineering since the very beginning, even when we were just in the US and Canada.)
  • In contrast, Flipkart has just one engineering center in Bangalore for its 1000+ person engineering team, comprised almost entirely of people from India.

The cost structure also varies per market, where some local engineering salaries can come in at a fraction (¼ — ⅕) of US averages. Many of the development practices I’m familiar with are built on the assumption that engineering time is scarce and human labor is costly— it makes sense to optimize for effectiveness and invest in robust systems/tools that can operate without human intervention. Discarding this assumption can lead to entirely different development norms.

One engineering leader told me about a team that skipped code reviews entirely, since it was faster to keep building and just fix forward as issues came up in production. A friend who worked for a Chinese startup divvied up work by handing higher-level objectives to many tech leads in parallel, while each lead fully directed the work for 2 much less experienced engineers in the background — interacting with each cluster of 3 was roughly equivalent to working with one engineer in the US.

Balanced, agile development vs top-down tiering (Scrum Sprint 1 image courtesy of Atlassian)

If you’re operating with a balanced 6–8 person engineering team using an iterative development style, you might painstakingly divide each project up into JIRA sprint tasks and manage execution collaboratively. If you have 3x as many engineers with a clear tiering of skills, you might get just as much done using a simple whiteboard and top-down assignment of handwritten tasks. Similarly, if manual labor is costly and inefficient, you might invest engineering time in building or integrating tools for automation. If that cost is significantly reduced, you’re probably better off keeping your engineers focused on product development.

Note: In many parts of Asia, the human labor vs. automation tradeoff can be about effectiveness rather than cost. In a world where digital infrastructure is still lacking and data speeds are slow, your online marketing dollars may go a longer way if spent on people knocking on doors. Uber and Lyft streamline mobile onboarding flows and buy ads to acquire customers; GO-JEK books national baseball stadiums to register thousands of drivers at once.

Engineering management is nascent and needed

Given the momentum of startups and investment in the region, there’s a huge demand for experienced engineering management. This is particularly acute in Southeast Asia, where few companies have seen enough scale (yet) to have developed deep talent internally, even as they enter hyper-growth and race to add engineers at a tremendous pace. In India, management culture is more focused on delivery and execution, while long-tenured engineering leaders continue to contribute technically through code or architectural design.

New managers are stretched to support teams larger than any they’ve been on before, and finding resources to develop them is a hot topic amongst startup founders. IC engineers feel this too, and many people I’ve interviewed list wanting to work with experienced mentors and managers as a top reason for looking for a new job. As a result, companies commonly look to other markets to hire in senior leadership.

Scaling engineering organizations is something I care a lot about — I’m curious to meet other leaders in the region and see if there are ways we can bridge the gap and bring in best practices, beyond importing talent. This also feels like a potential arbitrage opportunity for those in the mentorship business: there are engineering management circles/meetups/tech talks happening weekly in SF, while I’ve yet to hear of any sustained programs here in Singapore.

What else?

What have you noticed about engineering management and culture in other parts of the world? I’d love to hear your requests for topics to write about, recommendations for people to meet, or observations of your own.

Thanks to Gloria Lin, Dian Rosanti, Edward Chiang, Michael Manapat and Sara Tillim Adler for their help with this series of posts.

Next up: I’ll share some things I learned about India’s innovative approach to digital infrastructure.