Working with developers, startups to build cloud-native apps: IBM’s Seema Kumar

Tech giant IBM Corp has been battling Amazon Web Services, Microsoft and Google to grab a bigger pie of cloud services revenue in India. The company has been training startups and helping enterprise developers in providing cloud solutions. In an interview with TechCircle, Seema Kumar, country leader for developer ecosystems and startups at IBM India and South Asia, explains the company's strategy to rope in startups and developers. Excerpts:

What is IBM’s overview of the developer and startups ecosystem?

We look at it as one ecosystem as a whole – developers and startups. A developer can be a startup developer or sitting in a large enterprise or smaller firms. This makes the developer a very important influencer in the way technology is being adopted and consumed.

This is more so because there is a shift in the way decision-making happens today, from a top-down to a bottom-up approach. Today, it depends more on what technology your development team is using and what cloud platforms they are comfortable with to base your decisions on.

When I look at the startup space broadly, it includes early-stage startups to matured-stage startups and they are very next-generation from a technology perspective. 

The appetite for cloud technology, artificial intelligence (AI) technology is far higher in these new-age tech companies than traditional companies. The traditional companies are battling legacy issues, they do not want to disturb the infrastructure they have.

What is your strategy to help developers?

Our approach is tailored to cater to their requirements. For an enterprise developer, there are multiple paths in terms of journey to cloud. It can be a hybrid approach, where you are trying to extend your existing applications into the cloud in a secure manner or it could be a path under which you are building newer applications and you want it in a cloud native format, where it is based on micro-services, server-less architecture.

Our approach to enterprise developers is focussed on these typical adoption patterns or pathways to the cloud. What we also do is equip and arm these developers with a lot of content around that. One thing is the code, where you can build the code but you also need the right level of skills, documentation and so on and so forth to build the patterns and application. Recently, we announced a series of IBM code patterns digitally.

Any developer can go there and pick up a pattern of their choice. These are built on technology like Kubernetes-based containers, Internet of Things-based solutions and blockchain. One can pick these up, infuse into their present solution, customise as per their requirement and get on running with it. This is what we are doing largely from an enterprise developer perspective and enabling them with newer technologies.

What is your strategy for startups?

With startups, the approach is totally different. When you talk about technology startups, they are already ingrained with technology and you don’t need to do too much of an enablement or coaching. The effort is to help them understand, the benefits of basing their solution on IBM platforms and help them scale, match up with security requirements, performance and availability requirements etc. 

Any startup can apply for Global Entrepreneur Program (GEP) where they can choose among three levels depending on their journey and growth plans. They get access to credits on IBM Cloud. They get anywhere between $1,000 and $10,000 per month of cloud credits.

How do you decide which startups to help? Could you explain the benefits of GEP?

There is an application process. There are three levels. For the basic level, any startup can apply. This is open for all and there is no approval process. You get an account on the cloud with minimal functionalities.

Once you have familiarized yourself with what is available on the platform and when you are ready to build a prototype, you can upgrade to a builder account. For the builder account, a startup should be less than five years old and shouldn't have had less than $1 million of revenue in the last 12 months. This is where you get a credit of $1,000.

Then we upgrade you to the premium level, which gives you around $10,000 worth of credits. This step has an approval process. We look for somebody backed by one of the accelerators or venture capital firms or there needs to be some recommendation.

Which accelerators and VCs are you working with in India?

In India, we work with pretty much most of them. We work with Nasscom, T-hub in Hyderabad and some known startups in Mumbai. We do events together and work together in various ways so that startups not only adopt the platform but also grow with that. So we also help them in terms of access to clients and help them build their business.

For developers, we do a lot of community-based activities... We do a lot of meet-ups, a lot of hands-on workshops for free. These are Watson-in-a-day, IoT-in-a-day, Blockchain-in-a-day sort of workshops where developers sign up to come spend a day with the IBM’s team and learn some hands-on coding and these technologies. That has proved very rewarding for the developers as it helps them upgrade their skills.

Another major activity we have for developers in India is hackathon. Some customers run corporate hackathons – focussed on solving a few specific problem areas. 

We collaborate with a lot of our clients to help them tap into the innovation that is out there. We help clients bring in that innovation and make it real not just by hackathons but even after that – in terms of doing a proof-of-concept, running an application, using a new kind of technology which hasn’t been tested before.

Could you share examples of such innovations?

One example of a startup which came to us through our developer programme is Superfan.ai. They are into the business of building conversational interfaces primarily from a fan engagement standpoint. They are making a good use of the Watson conversation services APIs to build these chatbots.

When you look at the cloud as a platform, the whole idea is about driving next-generation workloads, talking about huge amounts of data where there is a requirement to do AI, machine learning (ML), deep-learning kind of algorithms. You need additional capacity in terms of GPUs (graphics processing units), which we offer on the IBM cloud. 

SigTuple is one of the startups that came through our IBM SmartCamp initiative. They digitise your clinical samples. For e.g., your blood test is digitised. If you are in a remote area where there are no qualified pathologists, all you need is a small centre with their device where the blood samples can be collected and the entire digitised sample is uploaded to the cloud and they run their algorithms on the cloud to do the diagnosis and generate reports.

Since you work closely with both developers and startups what are the new solutions you are looking at?

It’s more about what are the new problems we are trying to solve and the problems today are a lot around building for the cloud. In the past, the applications were not cloud-native. Unless your applications are designed and architected to leverage some of the cloud benefits, you are not going to realise their potential even if you have them on the cloud. So we are looking at a cloud-native architecture.

We are looking at a lot of developers taking up skills on container, Kubernetes, micro services, on server-less computing to build applications for the future.

When you move away from the core cloud, underlying containers and infrastructure to the application layer, every other application is going to be AI-enabled. There is going to be some level of AI capability embedded into an application, be it language understanding, visual understanding or understanding unstructured data.

In order to build that capability into your application where it is intelligent, it doesn’t mean you necessarily need to be very deep data science engineers. There are a lot of APIs available to do the basic AI-enabled tasks and that’s where Watson APIs come into the picture. At the same time, you will have use cases where there is a lot of custom AI-based solutions to be built – especially when you are dealing with hordes of unstructured data.

There will be a need to create ML models out of the data and training the right sets of data for different domains. Within healthcare, there will be many domains in which you need to train the models. All those are immense opportunities for developers.

---