AI Libraries for Developers – Hemanth Manda, IBM

This is Textron TV. Hey everyone, welcome back to another text on TV segment. I have some really I think it's gonna be an interesting interview great discussion.

I want to introduce you to hammont. Manda hamant is the head of strategic Partnerships for data and AI over at IBM a month. Welcome to text strong TV.

Thanks Allen. Thanks for having me. It's a pleasure to have you on so look, it sounds like a really cool job head of strategic Partnerships for data and Ai and IBM one.

Why don't we start there? They'll give people a little bit, you know beyond the title what your role is? Yeah, it's a very interesting role Allen previously.

I was in product management, but this is really cool. I get to work with Partners both on the inbound side of the spectrum where we are working with their products to take them through our go to market Channel, but more importantly taking our products through their go to market Channel, which is what we're going to talk more about it. Also I get to work with hyperscala such as AWS and is your so it's looking at inbound outbound as well as hyperscal Marketplace.

Excellent now. You IBM and your team made an announcement. I guess it was back in late, October.

you know IBM had historically has really helped seed so many. Different areas in technology, right? I I don't think they get a lot of credit for all the code.

They contributed to Linux for instance. Right? And and I mean name name an area I though usually significant IBM can contributions around there somewhere.

And is very much true in the area of AI. And you know everything from Watson, which is I guess the for many of us. The first time AI was real right was when we started seeing some of the things Watson was able to do but recently IBM made an announcement again helping to see this whole industry When I don't want to steal your story the one that you tell us, yeah, absolutely Anna.

I mean this is actually very exciting announcement and I feel it's gonna resonate with a lot of your viewers historically IBM, right, especially on the area of AI we were focused on applications and solutions looking at and outcomes, right whether it is industry specifics such as what's in health or horizontal Solutions around customer care. What's not sustained what's in Discovery extra. That's what we were historically focused on but what we announced in October is what we calling is embeddably AI which is focused on infusing our foundational AI Technologies coming from Watson into applications Solutions made by other software vendors or Enterprises, right think of this is more of an embed play and it's extremely critical because I feel like world is moving towards embracing AI everyone is building applications and they want to infuse Ai and they have a choice either they build this on their own or partner with the likes of IBM to essentially take our foundational Technologies and included within their applications.

So that's that's the gist of what we announced and we're seeing a huge amount of demand around it. I would imagine. What I want to discuss a couple aspects of this but let's start with you know, there were three libraries that are initially made available.

If you would why don't we just dive into not a deep dive because we don't have that much time but a dive into each of the three of them talk about use cases around them. Yeah, absolutely. I think the the three libraries we announced were natural language processing speech to text and text to speech and and I think all of them are I believe of fundamental to most of the AI automation that's happening in applications installations today and we can yeah, we can walk through use cases around all three of them.

Give us an okay. Go ahead. Give us a few use cases if you wouldn't mind.

Yeah, absolutely. So on natural language processing, right? This is about essentially understanding the human language whether it is sentiment analysis whether it is basically figuring out what the discussion is about summarization categorization Etc.

So to dumb it down, right like one of the use cases we're working with an ISP. I cannot name the isv because we still in initial discussions here think about documents, right? It's think about specifically contracts what typically happens is you have personal identifiable information in the contracts such as your name address date of birth social security number Etc.

And when you're sharing with Partners you want to identify that pii information and automatically reacted right now, it's all manual. So you have to manually go through these contracts identify what your Pi information is and use a use a pen to Clearly redact them before you start sharing them. It's extremely manual time consuming.

What this ISP is doing is working with IBM leveraging our NLP technology what the click of a button they want to identify the pin information and redact it. Obviously, there's some training that happens and the training is infused into the in the into the solution itself, but they're looking at leveraging AI to automate it. So this is one example use case that I can talk about.

Obviously. There's a lot more Absolutely. You know what?

Let's get this out of the way now for people who want to dive deeper. Into the use cases into these specific libraries. We'll try to get the exact URL in our notes.

But just quickly where you know, where can they where do we start on the IBM site to to find that info? Yeah. So these are available on our website.

There is specific website for natural language processing and speech Are both the detail use cases also trial a six-month trial if they want to try it out themselves. And we also have an experience our self-service trial which is you don't need to deploy anything. If you just want to try out the art of the possible.

We are also making it available. So all of these available on our website Okay, and we we have the URL. We'll try to stick that in the notes there.

so Beat your plans. I want to talk, you know a good part of our audience is the developer audience right software developers and devops people Cloud native folks. want to make clear that these libraries they're not in and of themselves applications their libraries.

Hence the name libraries, right? That developers can incorporate. Into their applications.

So this is really, you know, part of the developer tools if you will which you know, look IBM's good part of their bread and butter was always giving developers the tools. They need right to be successful. Let's talk about how does it developer?

Use these libraries, right? How would we build applications with them? Yeah, it's a very good question.

The one of the things that I would like to emphasize is this particular launch right around ambitably AI it is very much focused on developers because we want to bring AI to developers and part of that focus is to essentially ensure whatever we are shipping has apis has documentation. It's developer friendly and we also have flexibility right because developers would like to basically pick and choose what is best for their specific application or use case. So so we have the way we pack is the various structured.

This is by ensuring this is developable friendly and not only that we are ensuring that the framework is consistent across our libraries. So if you want if you're the developer would like to include like more than happy and speech or in the future and I'll be in Time series The the framework and the AP. Structure would remain the same so that it's easy for you to include them.

You don't need to go through a learning process or training exercise to understand how to include one Library, which is the other Actually, excellent. All right. so You know in talking off camera these three libraries that IBM is now made available are really just kind of the tip of the iceberg.

Yeah, absolutely. I think we have we have significant plans for future and we want to expand the the set of libraries we offer. Sure.

So without I don't want to get you in trouble or me in trouble. But any can you give us any insight into kind of time frames for future libraries and and you know what these libraries may be about. Yeah, absolutely Alan.

So next year. We're applying to launch at least two or three more libraries things such as document understanding which includes OCR and other capabilities to be able to be able to read and understand documents. We are also looking at time series, which is around forecasting right?

There's a lot of applications out there which want to predict the future and they typically leverage time series libraries. So that's something that we applying to launch. There's also a lot of focus around Watson trust libraries because the thing about AI is while it is great, you can automate and credit future but being responsible being ethical and ensuring that trust is infused within AI becomes extremely critical especially as Enterprises and software Windows speaking to embrace and scale AI Also, we have a set of libraries around trust that we also would like to commercialize next year or the year after.

Sure. You know. Uh, this is so much hype there's so much Buzz.

Around AI I will tell you even an hour little business here. you know every day our editorial team is messing around with these AI article generators what you put in a couple of ideas and it it writes the story for you, right and and we're watching it closely you could still Separate an AI generated story from a human for the most part from a human written story. It's not quite there, but it's getting there.

There's no doubt about it last week. I saw another one a PR an AI empowered PR agency. So this PR agency just uses AI to pitch your stories and publisher releases and do everything kind of like Robo PR.

So there are there's real world examples, of course the whole Watson. Saying is is real right, but there's also so much. Overhyping buzz and future talk around AI.

Where does this fit on that? You know, I look it's IBM you guys can't be? Bsing it, right?

Rubber meets the road here. Have we had people develop applications with these libraries already? If so, I can you point to any of them.

Yeah, absolutely. So it's funny that you ask because the Genesis of this whole initiative is internal. We being using like we have 20 plus products that currently include an embed natural language processing.

So we've been using this internally for the past few years and we saw a lot of benefit our developers, right our internal developers a lot of benefit from these libraries and that's when IBM decided the reason here to basically make this available to external customers as well who want to obviously leverage this the same way maybe in love waiting for the past few years. So this is not us we've been using it we have tested out a lot of our products have been used by thousands of customers that embed these libraries and all we're doing right now is to make these available externally so that our customers and our partners can also benefit. I love it good stuff.

Amanda I think I've got all the questions were about out of time. Anyways there anything else you wanted to add though? Um, I just want to I just want to go back on your observation, which is a very valid observation.

That doesn't but a lot of us around AI to me AI is three things and three things only. It's either about predicting the future which is around forecasting or about automating things right typically mundane things, but you can also automate complex things if you put enough effort into it and third is around optimization. So anything about and Beyond those three in my opinion is pure Buzz forecasting Automation and optimization.

So look at any use case. I think it would fall into one of those three buckets. That's so great advice for them again order Nation.

forecasting optimization Yeah, those are the three things where AI is real today, and we should be looking at. Excellent advice hermit with this great news from IBM. We're good.

I'm can't wait to see the next libraries coming out and and I can't wait to see people actually using these libraries and incorporating them into their application. So it'll be exciting. Congratulations to you in the whole team there.

Keep us posted. Thank you, Alan, and thanks for having me my pleasure. We're gonna take a break here on Tech strong TV.

We'll be back in just a minute.