Building The Defences That Protect Millions Of Indians Every Single Day—Praveen Ravula’s Solutions

An AWS Security expert averted major disruptions and built high-speed threat detection systems, ensuring India’s digital services are always reliable, secure, and available. 

India’s digital economy is growing at a phenomenal rate, with the IDC predicting the country’s public cloud services market could hit $30.4 billion by 2029. Yet, an insidious threat looms: even though all the country’s critical digital services virtually depend on very large cloud computing systems like Amazon Web Services (AWS) for their speed, safety, and accuracy, hackers are constantly targeting them to steal people’s money and information. Imagine a situation where such a massive system faces a security problem or stops working; the effects would be enormous: payments would fail, online stores would close, and millions would suffer.

So, when the scale of such problems becomes this large, it will be almost impossible for humans to keep up with it. Hence, there is an urgent need to build self-healing systems that can always detect and mitigate such threats—a challenge Praveen Ravula tackles daily as a software development engineer for AWS Security at Amazon. Praveen was educated at the University of Florida for his master’s degrees in computer science and is certified in AWS Academy Cloud Foundations and Microsoft Azure AI Fundamentals. He previously worked as a software engineer at SDK Tech. Thereafter, he joined Nebula Partners as a full-stack web developer. Now, he leverages those experiences to build systems that fortify AWS against cyberattacks. So far, his systems have demonstrated a high threat detection rate in just seconds and reduced AWS data transfer costs considerably, making its cloud services highly secure, faster, cheaper, and readily available to millions who rely on them daily. Yet, a few important questions remain: since cyberattacks are inevitable, how can engineers build self-healing systems that can block these attacks and save costs at the same time? What lessons can Indian companies learn from Praveen’s engineering exploits as they protect their fast-growing digital future?

Praveen, you led the Dogfish Dataset Regionalisation project at AWS, significantly reducing the cost of data transfer. What security need prompted that massive data redesign, and how did you achieve such an impressive result?

At AWS scale, security, performance, and cost are deeply interwoven. The Dogfish dataset is very important to us because it provides all the information our automated systems need to make accurate and timely security decisions. However, we had a major security problem at one point because our old system depended on data from different regions. This dependence slowed down our Dogfish Enricher Lambdas and increased security risks. So, we had to find a lasting solution: migrate the architecture to a regional S3 bucket design. This minimised delays in our security services. Also, it cut the cost of data transfer by a quarter, thus reducing our operational expenses. Now, our Indian customers enjoy faster and cheaper services. This means that consumers across the country experience quicker transaction approvals and better service delivery.

Besides, you have built custom systems like the Allowlist Workflow with MadPot detection, mitigating threats with high accuracy. What is the underlying risk of using generic security solutions at AWS scale, and how does your TypeScript-based workflow explicitly protect the integrity of the services consumed by Indians and others?

Generic tools attempt to fit all situations, but AWS requires a high degree of precision. So, we are duty-bound to protect millions of customers without blocking legitimate users. A generic tool may create too many false signals, thus cutting off vital systems or hindering genuine banking operations in our country. That is why custom systems have to be built. The Allowlist Workflow uses tools like MadPot detection and special settings to ensure that we make impeccable IP allowlisting decisions. This degree of accuracy ensures that users in our country and elsewhere are protected from hackers, who might want to steal their vital information or misuse our services. In addition, I modernised an existing  Allowlist Policy into a TypeScript-based workflow with 100% unit-test coverage. This ensures that the system’s logic is always verifiable, maintained, and reliable as the cloud environment keeps changing.

You need a certain interdisciplinary skill set to modernise such a system and make it efficient. How did your previous roles at SDK Tech and Nebula Partners prepare you to confront such a huge challenge head-on?

From experience, I knew that resilience is something you build through engineering. So, my previous experiences at SDK Tech and Nebula Partners helped me understand how things actually work in real life. There, I optimised CI/CD pipelines and Docker container setups, reducing deployment downtimes by 90%. Currently at AWS, I quickly realised that the same principle of high availability applies everywhere. So, the same approach for a safe rollout of a new feature also applies to fixing a security patch immediately. 

Your recently co-authored research, published on IEEE Xplore, discussed the use of machine learning to detect VoIP spam. In particular, it utilised high-performance computing (HPC) in order to achieve a high detection rate under a minute. How did the data modelling and high-speed methodology from your research inform your strategy for reducing large-scale security issues at AWS?

To be honest with you, the study, which I carried out with other researchers, made me understand that it is possible to use smart systems to solve issues in real-time. In that project, we developed a model to differentiate fake calls from real ones. To do that, we analysed user behaviour using directed, weighted call graphs. In addition, we used HPC to process very large amounts of data quickly. In the end, we were able to reach 98% detection within just half a minute. Now, such a mindset—diagnosing an abnormal behaviour and remedying it immediately—is the basis of my work at AWS. It alerts our systems, such as the Mitigation Publisher Service, to make security decisions automatically. That notwithstanding, I understood the importance of speed, especially at the time I handled a Severity 2 ticket during the October 2025 AWS outage. During a global emergency like this, you probably have no time to start writing new code. Instead, you fall back on smart systems that have been previously built to take action within seconds, not even minutes. In a nutshell, the study gave me the strategy to build such high-velocity automated defences.

You have been elected as a Fellow of Hackathon Raptors, an international technology organization that brings together leading engineers from more than 30 countries. How do collaborations within global professional communities help engineers stay ahead? 

That is a good question. Technology changes too quickly for engineers to grow in isolation anymore. Communities like Hackathon Raptors foster environments where professionals from different countries and industries can exchange experiences and learn how others solve complex technical problems. I think these collaborations are also valuable because they expose you to new perspectives and practical approaches that you may never encounter within a single company. They also encourage continuous learning and help engineers stay adaptable, which is especially important in cloud security.

Praveen, you are rightly considered one of the top professionals in cloud security and large-scale infrastructure engineering. How does your Instagram page with over 150,000 followers help inspire and train the next generation of cloud security engineers?

The lack of expertise in cloud security is one of the major problems affecting our digital future in this country. To address this problem, I created a platform where I always share free, valuable tips on cloud security. To date, many of my followers have been able to learn real engineering skills through my tutorials. That is why I am glad about the progress we have made within just one year of setting up the platform. 

However, I would advise our future engineers to master the engineering principles of resilience instead of just learning security tools. That would make them experts in the field. Then, for our companies, they should never see security as an afterthought but build it into their systems from day one. Finally, they should start automating compliance right away because a stitch in time saves nine. Otherwise, they risk losing billions of dollars in fines for simply failing to protect themselves and their customers.

Published on: Tuesday, August 04, 2026, 03:21 PM IST