Democratising ai by making it available to all developers amazon’s manaktala techcircle data recovery disk

Amazon Web Services (AWS), the cloud computing arm of Seattle-headquartered tech giant Amazon, has been locked in a tussle with rivals Microsoft and Google for market share both in India and globally as more companies look to shift their operations to the cloud.

In a wide-ranging interview with TechCircle, Navdeep Manaktala, the head of business development for Amazon Internet Services Pvt. Ltd, said that the company aims to democratise AI and make it available to the entire developer community. Edited Excerpts:

We sort of categorised our portfolio of AI services into three layers – at the top is the API (application programme interface) layer, followed by a mid-tier, and at the bottom is the most advanced layer that we designed for data scientists or people who are comfortable and familiar with training huge data models, testing them and then fine-tuning those models.

This is where the Amazon Sagemaker platform comes in. It is a fully- managed machine learning service that provides an integrated Jupyter authoring notebook instance for easy access to your data sources for exploration and analysis, so you don’t have to manage servers.

If you go into AI and ML overall, there are two things that you need. One is that you need tonnes of data to be able to train your models. This is a big challenge for common developers as they don’t have access to any such data sets. The second issue for them is to get access to huge computational power, which again is missing.

We offer the same computational power required through our cloud business and this includes multiple flavours of GPUs (graphics processing unit). So if you put all of this together, what we are really doing is democratising AI by making sure that these services are generally available to the entire developer community.

AWS has a lot of customers in India. Haptik is one of the hottest AI startups in India that has been using our service. One of the major used features of Haptik is the reminder service. Earlier, they were pushing reminders as notifications which users might end up missing especially with too many notifications coming in these days. However, with the use of AWS Polly – a paid service that can convert text to speech in 47 voices across 24 languages – Haptik can now call its customers for reminders set on the app.

PolicyBazar has used text-to-speech and that pretty much enables them to change the message in a minute, as opposed to standard IVR which is changed once every six months or so. For example, they can run a promotion for an event any time of day or run a different voice and different language to a different customer. All kinds of customisations are possible.

At company, a lot of employees used to do grunt work in terms of identifying pictures uploaded by users. Then they started using Amazon Rekognition. The service makes it easy to add image and video analysis to your applications. The service can identify objects, people, faces, text, scenes, and activities, as well as detect any inappropriate content.

The first part of the problem is to be able to aggregate that data from multiple sources. If you look at any big enterprise today, I think the biggest challenge for them is that despite having data sources (from a Customer Resource Management or Enterprise Resource Planning system or clickstream data from a mobile or web app), they sit at different silos within the organisation making it difficult to take large actions.

The other material trend we are seeing and enabling for customers is to be able to aggregate all the data that your organisation is generating, essentially forming a data lake. If you look at Amazon S3, which is a searchable storage service, it becomes the single source of truth for enterprises with multiple uses for departments across business intelligence or data scientists and other business users who are firing sequel queries.

We also provide some tools for these searches as well such as Amazon Redshift, which is a data warehousing service, and Athena, an interactive query service that makes it easy to analyse big data in S3 using standard SQL. We even have tools to catalogue the data set irrespective of the data format by using Amazon Glue.

This is not just a proof of concept or prototype that we are building. Businesses today are powered by our AI and ML solutions and we are making those same pieces of technology available to the general developers community. I think that is one of the major differentiators.

We also conduct online training, in-account trainings for large accounts as well as training sessions at our offices which customers and partners can attend. We also have a whole set of certifications and accreditations which are becoming increasingly popular.

We are also talking about how to attack specific use cases surrounding AI. We have just announced a ML competency programme under which a partner gets a certification in Amazon’s ML solutions and then that partner can help customers in devising solutions around AI.