Infobip’s Tuzla office opened in 2019 as the company’s second office in Bosnia and Herzegovina. Today, it’s home to the engineers who own Infobip’s email product stack end to end, building and operating systems that help companies reach people around the world.
The scale is visible in the numbers: the platform handles around a billion emails a day, with spikes of roughly 35,000 emails per second. During peak periods such as Thanksgiving and Black Friday, the team prepares for billions of emails over a few days. But behind those numbers are the decisions, teamwork and day-to-day engineering that keep the platform moving.
“We’re building and operating a global email platform, handling around a billion emails a day, from right here.”
— Ermin Katardžić, Principal Engineer
Five years ago, Ermin was a mid-level engineer. Today, he’s a Principal Engineer, helping shape the technical direction of the email platform.
When the Tuzla team took on more responsibility for email, the platform looked very different. Incidents were frequent, much of the traffic went through a third-party provider, and Infobip’s own mail transfer agent (MTA), the software that sends emails, handled a few hundred thousand messages a day. The team focused first on reducing incidents, then on improving the platform and moving more traffic onto Infobip’s own infrastructure.
It also changed the team’s role. Tuzla engineers now contribute to the technical decisions that shape how the platform grows and stays reliable.
“We don’t just build pieces of someone else’s system, we own the engineering end to end, from design to production.”
— Ermin Katardžić, Principal Engineer
As Ermin has grown in his role, he’s developed a broader understanding of the platform. As a Principal Engineer, his focus shifts with the work: sometimes he codes every day for weeks; at other times, he contributes through design discussions, code reviews or investigating production issues. He stays close to the highest-priority work, while helping the team make decisions that will hold up over time.
One lesson came from a feature intended to control which sending IP addresses could reach particular destinations. It started as a practical fix, but its allowlist design meant that newly provisioned IPs could be excluded unless someone updated the configuration. The team spotted the issue while reviewing how new IPs were warming up: they weren’t sending traffic to iCloud, even though the infrastructure was ready.
The feature was eventually removed. For Ermin, it was a reminder that even a temporary solution can create a lasting operational cost, and that a product team needs to think about how decisions will affect the platform long after they’re first implemented.
Peak season isn’t something the email team leaves to chance. Preparation starts well in advance with performance stress tests and conversations with customers about their expected traffic.
Those projections help the team plan capacity and warm up additional sending IPs. That process takes time: traffic needs to increase gradually before an IP can handle a significantly higher volume. The team also scales services and watches the MTA closely, where CPU and outbound network capacity are among the key pressure points.
Ermin has also helped shape how the platform handles outages. For outbound email in the US and EU, he advocated for an active-active setup: IPs in different locations are both actively sending traffic, so one location can take on more if another goes down. The trade-off is that traffic through the surviving IPs may double, which can lead to throttling by mailbox providers. An alternative could reduce that risk but involve downtime while traffic is moved.
There isn’t a perfect option for every situation. The team has to understand the platform, weigh the consequences and choose an approach that works for customers in production.
Maida Baralić joined Infobip as a Senior Engineer after hearing about an interesting project coming to Tuzla. Four years later, she’s a Staff Engineer. Much of her role involves keeping the platform reliable and thinking through how future features should fit into the systems already in place.
One example started with a caching layer for templated HTML. Caching is useful when content can be reused. But one client was sending unique HTML for every message, and that content was being cached even though it would never be used again. The unused entries were taking up space and pushing reusable content out of the cache.
Maida traced the issue to that client’s traffic and added its account to an ignore list, so its HTML would skip the cache. The change also reduced internal network usage and made overall processing faster.
“We process billions of messages every month, so performance under that load is something we always have to think about, and there’s plenty of room for optimization. You get to work on systems used by companies around the world without leaving Tuzla.”
— Maida Baralić, Staff Engineer
Much of Maida’s most challenging work happens behind the scenes: tracking changes in sender reputation, authentication and mailbox-provider rules, and catching unusual patterns before they affect customers. With large clients, a small issue can affect a huge number of messages, so preventing problems early is part of keeping the platform dependable.
“When we do that well, it looks like nothing happened, and that’s exactly the goal.”
— Maida Baralić, Staff Engineer
Maida has also seen the Tuzla hub grow in both team size and scope. Over the years, both the team and the platform have grown, taking on more work and larger clients. She says the atmosphere has remained friendly and collaborative, even as the work has become more demanding.
Not everyone joins the team with years of experience in email infrastructure. For Mak, the Tuzla hub became a place to learn an entirely new domain while contributing to a global product.
His first production incident came early. After a code change went live in Frankfurt, around 4,500 emails were rejected without a retry. Mak spotted the errors and disabled the affected instances. The incident was a learning experience, but what stood out to Mak was the team’s supportive response.
“The team was incredibly supportive and encouraging. They immediately shared stories of their own first incidents and kept reassuring me that ‘now I’m finally properly onboarded ‘.’”
— Mak Redžić, Software Engineer
Soon after onboarding, Mak was given the opportunity to redesign how the MTA handles its thousands of public-facing IP addresses. It was important work, but not urgent, giving him time to think through the solution and show what he could do.
Recently promoted to mid-level engineer, Mak says that growing domain knowledge has brought a greater sense of responsibility. He’s also learned by watching more experienced engineers work through problems and connect the different parts of the platform.
Working in a product company has given him experience across the product lifecycle, from planning to implementation, alongside colleagues in product and engineering. For candidates, that means the work doesn’t stop at shipping a change: the team stays connected to how the product performs in the real world.
“Working on email in Tuzla means working and growing as an engineer on a global scale.”
— Mak Redžić, Software Engineer
The stories from Ermin, Maida and Mak show different ways of growing at the hub: shaping the architecture of a mature platform, finding a small change with a measurable effect, or taking on real responsibility early in a career.
Together, they point to what Tuzla offers engineers: a chance to work on a product used around the world, contribute to the systems behind it, and learn alongside people who understand it deeply, all while building a career in Bosnia and Herzegovina.
