Behind the Code: Engineering Year in Review 2024

Reading Time: 6 minutes

I want to start the 2024 year review with the mindset that has always worked for us and moved MoEngage engineering in the direct direction.

 I have seen the best engineers have this mindset intrinsically. This mindset is needed for any hungry organization like us. We always believed we had a lot of room to improve, but improvement needs to be driven by every individual in the team, and that’s when we can achieve our maximum potential. I noticed MoEngage Engineering diverging slightly from this mindset, and one of the significant efforts in the latter half of 2024 was to bring this mindset back into every team member.

Now that you know what we believe in, I wanted to share some of the significant changes we made in Engineering in 2024.

Vision

Starting in 2024, we formalized our vision for MoEngage Engineering. This vision statement has guided us in making decisions that are streamlined and easy to make. I am thrilled that we have a guiding post for our team members.

Scale

The scale at which our platform operates has challenged our team for years, and our challenges grew by about 30% last year. It’s easy to build a simple application and make it work in production, but can you tune it to balance SLAs and costs(customer love and business P&L) for 2 billion monthly active users? That’s where the secret of the sauce lies, and the challenge goes to the next level. We are fortunate to have these challenges daily and solve them actively.

We launched a new data center for our customers in Indonesia, taking the total count of our worldwide data centers to 6. In 2025, we are cooking a few more recipes to increase our global footprint and expand our presence across newer countries.

Security

In 2024, we tried to shift left as much as possible in securing our data. We have mandated a few of the best practices for our team, and secure design review has become integral to our development process. We have improved our DLP practices across the software applications used by our teams. We have improved our cloud security posture, allowing our team members to access the data they have access to securely. Our team has also tried onboarding GenAI products into the operations to save time, something to look forward to in 2025.

Due to some unforeseen circumstances, we have fallen a bit short on our goals for 2024, but our security team is going full steam ahead into 2025 with a lot of momentum and the right targets to achieve.

QA

We all know bugs and incidents are unavoidable in production. Over the last few years, we have made significant investments to shift the testing leftwards and become more reliable. Naturally, as part of shifting left, we have set a goal of improving automation coverage for all products in 2024; we are headed in the right direction here. We saved 105 testing hours per release cycle, which is huge.

Apart from shifting left, we have set a new goal to identify these anomalies in production before they affect most of our customers and proactively fix them to limit the damage to our customers’ experience.

We have launched a concept called Blackbox monitoring, which runs a suite of test cases in production every X hours, and if any cases fail, we will identify the cause and fix them. I am proud to say that we have averted around 40+ issues from blowing up in 2024. Our QA team has embraced this project wholeheartedly, and we can see its fruits.

SDKs

How do we automate our releases for various maven repositories out there? The SDK team has improved the world by contributing our solution to one of the problems we faced. Our contributions have been listed in the community plugins on the Maven central repository page here

The SDK team also embraced the Kotlin multi-platform approach, where the same code can be used for iOS and Android. This reduces our development time and brings consistency across the platforms, as SDKs are the first touch points of integration for our customers.

Frontend

Finally, we removed Angular code from our code base, significantly boosting our development times and testing cycles. MoEngage frontend code is entirely in React now. We took on this daunting mission three years ago but have committed and achieved it.

One innovative thing the FE team has achieved is moving to micro frontends, where each team can maintain its code base and still work as a single app from the customer perspective. Most of the FE teams at MoEngage have moved to micro frontends, and on average, each team is experiencing a 3-5 day gain in release times. This is a fantastic effort from our team, and I couldn’t have been happier.

Backend

I think we could have done a lot better here than we did in 2024. Nevertheless, we delivered some interesting outcomes to our customers and the business.

We have revamped one of the most coveted products of MoEngage – Flows. The new stack rollout is ongoing for some customers. This change will improve the reliability of Flows for our customers. Firebase Cloud Messaging has moved away from accepting multiple tokens in a single API to send push notifications. This was a massive surprise to us, but our team adopted it brilliantly, and HTTP2 multiplexing was introduced to counter this. We saved about $7k/month due to this change. We have also introduced FastAPI to one of our public-facing services, which receives burst traffic. This has reduced our p99 latencies for this service by about 10 times.

MoEngage data ingestion has become close to real-time, and we have greatly improved our customer experience. You can read more about that here. 93% of the time, we can provide data to our customers within 5 seconds of the data hitting our servers. The extension of this project is real-time segmentation for our customers, where we can help them interact with their audience accurately. Real-time segmentation capability has boosted our personalization experiences massively.

As part of our vision, we want to stress the platform services that will empower our application teams. We have launched or inaugurated a couple of services that will help our teams launch API or streaming services more easily than in the past. We have also done some experiments with introducing Iceberg tables into our framework but quickly realized that there are some fundamentals we need to fix before we get there. We will pursue this again in 2025.

Operations and Culture

One of my key observations was that we needed to raise the bar for engineering and hold ourselves accountable to higher standards. We have taken initial steps towards formalizing our performance evaluation process, which I see changing slightly in 2025. 

We also reboarded all our team members, emphasizing what we expect from each team member and what kind of support they can get from the leadership team.

Improvements

At the year’s end, I found plenty of areas where MoEngage engineering could be better. 

Operational excellence has been one of our Achilles’ heels, and we want to fix this in 2025. Hopefully, our plans will turn out well. We have outlined a clear action plan and targets we want to achieve in 2025. We will put our heads down and strive to achieve them.

Our platform teams could deliver better outcomes than what we achieved in 2025. I have always believed that a platform team can do wonders for any company, and it’s essential to get it right. We will keep fixing the problems identified until we do.

I also believe our teams can be more data-driven and impact-focused when constructing the OKRs—we intend to change that in 2025.

What’s in Store for 2025?

In 2025, we want to change how we look at our success. We have been looking at some micro metrics so far, but we want to change it to a macro metric, and we want this to align with our values as well. Customer obsession is one of the core values of MoEngage. With this cue, we will continue to measure the impact of our engineering efforts with a simple question –