Viral B. Shah, Co-Founder and CEO at Julia Computing chats with Shripati Acharya Managing Partner, Prime Venture Partners.

Listen to the podcast to learn about:

03:48 – Why build another programming language?

13:30 – Who was using Julia in the beginning?

17:10 – How an AI/Ml engineer can get started with Julia.

19:10 - How interested is Indian software community in Julia?

22:30 – Why Viral started Julia Computing.

25:00 - Open Source Business Model

30:49 - What applications entrepreneurs are building using Julia

33:02 - What is next for Julia and Julia Computing the company?

Read the complete transcript below

SHRIPATI ACHARYA 00:35

Hi and welcome to the Prime Ventures podcast. This is Shripati Acharya managing partner at prime. Our guest today is Viral B Shah. Viral is one of the Co-Creators of Julia, a programming language created and designed in 2009 specifically for high performance scientific computing. It is now used by over 10,000 companies and 1500 universities and 29 million downloads. Creators of Julia have won several prestigious awards, including the James H Wilkinson prize for numerical software. Viral is now co-founder and CEO of Julia computing and is now based in Boston. Welcome Viral to the Prime podcast.

VIRAL B SHAH 00:54

Thanks Shripati for that wonderful introduction. Really great to be here.

SHRIPATI ACHARYA 00:59

Absolutely. So let me start, rewind the clock a little bit and trace a bit of your personal journey for a moment. So tell us a bit about your journey from your college days, to your stay in India, to the creation of Julia, if you could actually draw that out for us that would be terrific.

VIRAL B SHAH 01:21

Yeah, so, I grew up in India, and I did my undergraduate studies. In the Mumbai University, Vasantdada Patil Pratishthan’s College in Mumbai. And I then went to University of California, Santa Barbara for a PhD. After that I worked for a company called Interactive Supercomputing, building parallel computing software. And that company at some point got acquired by Microsoft. And I decided that I wanted to sort of try something new and interesting, rather than sort of go into a traditional career path.

And I ended up as a result, moving back to India, which was one of the smartest decisions of my life. That’s when I met you for the first time at the Aadhaar project and I spent a few years building Aadhar and all the payment systems and all the foundations for what is now UPI and all the good stuff everyone knows about.

But I rolled out a lot of that infrastructure when I was at Aadhar along with Nandan Nilekani, we wrote a book called Rebooting India outlining all of our learnings of how to roll out big projects in government, of how to be a startup within government.

And the interesting thing is that when I moved back to India to start Aadhar, I started the Julia project at the same time, almost in the same week, along with three of my co-founders. Professor Alan Elmen, at MIT, and then Jeff Bezanson and Stefan Karpinski, all of whom are co founders of Julia computing today.

So in 2009, I started my work on Julia as well as Aadhar back then. Did Aadhar for a few years, and then Julia was taking off, it was becoming the language of choice for a number of different applications, and we started Julia computing. And that’s what I’ve been doing for the last six years.

SHRIPATI ACHARYA 03:40

Well, let’s double click on Julia a bit. So Julia, as we understand it, is a programming language. But there have been, the programming languages come and are created, like once a decade or every few decades, like Python dates back to 1991, or something like that when it was created. Then C, C++, probably created a decade earlier or two decades earlier in some cases. So these things just don’t come up that often. So what made you think that? Well, let me actually start thinking about creating a new language. I mean, that’s a fairly major undertaking.

VIRAL B SHAH 04:27

Yes. You know the interesting thing about Julia is that the four of us who came together to build Julia come from very different backgrounds, my personal background comes from sort of an applied mathematical, scientific computing background. So I really wanted to use computers to solve interesting scientific problems that have a real impact.

If you look at everything around us today, all the challenges we face whether it’s climate change, or whether it’s building a new electric car or whether it’s sending rockets to space or whether it’s developing a new vaccine. The innovation economy today is deeply scientific and deeply technical in nature.

And what I felt is at the start of my journey on Julia back in 2009, the tools that the computational scientist had, whether it’s a an engineer, whether it’s a data scientist, whether it’s a scientist, or whether it’s a quant, whether it’s an actuary, anyone who’s doing anything that is technical in nature, the tools they had at their disposal were primitive, and quite terrible, if I could say so. And the main drawback was what we call the two languages, that they all suffered from the two language problem that any technical user wants to write their code at the level of their algorithm. They want to sort of describe the mathematics or the algorithm at a higher level, and then press a button and hope that it’s just going to work at scale.

But in reality, that doesn’t happen. In reality, you use Python, or R or MATLAB, or SAS, if you are from an earlier generation, perhaps those are the kinds of languages people use to express their ideas in. And then you sort of take that code and throw it over the fence to a team of 50 or 100 programmers who will rewrite it in C or C++ or C# or Java. So you sort of had this dichotomy where you had the languages that were easy to use and languages that you used for production, and they were never the same.

And if you step back, it’s kind of funny to think I mean it’s obvious how bad this is. You’re writing everything twice, and achieving very little by redoing the work. And so this is what we call the two language problem. Julia started out as an answer to the two language problem. Can you have a language that is as easy as Python, and as fast as C? And if you sort of had this in one particular language, then it could just open up the doors for innovation.

I mean, now that I look back at it 10 years after having started Julia, it’s clear what innovation it could lead to, but back then it was just the germ of an idea that can you have performance and ease of use in the same language and until Julia came along, the answer was believed to be no, that it’s not possible like people always thought it’s almost like a law of nature that these two, you cannot have them together.

So that’s what sort of led to the creation of Julia. Just like everything else, I think it’s the itch. Like when you really feel the pain with something that you were using before you create something new to maybe try and address that.

SHRIPATI ACHARYA 07:50

So you were here in India, how did you link up with Alan Edelman and Stefen Karpinski, we’re in different parts of the globe, basically, to start brainstorming on this problem.

VISHAL B SHAH 08:08

That’s very interesting. Yeah. So it’s, they’re all sort of independent connections in my PhD career actually. And soon after. So Alan Edelman was someone who I met at MIT. My advisor, John Gilbert was visiting MIT, and took me along with him, which was one of the most fun times I ever had, spending a semester there and I met Alan Edelman back in 2003, for the first time.

Subsequently, he was on my thesis committee and everything and then Stefan Karpinski was actually a grad student in my department at Santa Barbara. And we knew each other for years because we used to run an ultimate frisbee team at Santa Barbara department, our ultimate frisbee team.

But we had never discussed research until it was time for me to leave Santa Barbara. And I was telling him about these horrible languages and problems in numerical computing. And he was, Yeah, I have the same, you know, views and same problems that I suffer from. And so that’s how I met Stefen back then.

And Jeff Bezanson is someone who I met at the first startup I worked at. He just walked into the door one day and said, Hey, I like working on programming languages, you have a job? And that’s how we joined that company when we first met. At some point when that company folded down and got acquired by Microsoft. Jeff and Alan and I said, let’s try something new.

And I was bouncing off this idea with Stephan as well at Ultimate Frisbee field. And we just did an email introduction, the four of us did an email conversation. Next week. After that conversation. Jeff wrote the first few lines of code. Two weeks later, we had the first working Julia language implementation, like something where you could do one plus one, and get something useful out of it. So we just did not spend a whole lot of time brainstorming, we just started writing code.

SHRIPATI ACHARYA 10:12

That’s fascinating, kind of like developing a language startup. Usually languages, especially high performance languages, I would reckon would be tied to hardware, isn’t it? I mean, did you actually start with some saying, hello, we need to actually start making it work on reference implementation on these kinds of hardware, and then you’ll actually go into other things.

VIRAL B SHAH 10:39

So we wanted to build something that was very portable, but if you think about it, these things are like, you can never foresee the future, right? Like, as you yourself pointed out that programming languages are probably one or two a decade come, that stick. And we’re still using Fortran from the 50s. And C is from the 70s. java is maybe late 90s, mid 90s kind of thing. So you kind of get like one or two per decade that really stick around.

And so I think you have to have a lot of hubris if you say that I’m starting a programming language, then, I have this immense roadmap, in my mind for what all I’m going to achieve with it. Only large companies have the luxury to do that. Sun microsystem was doing Java, they probably had some luxury along those lines. When Microsoft did Visual Basic or C# dotnet. They had probably that kind of roadmap and backing, financial backing to do that. We did not, so we just said we want it to be easy.

So dynamic programming language. So you don’t want to say what’s a double precision type, and what’s the single precision? And what’s the string? Because who has time for all that, that level of detail. And so that’s what the end user wants? And then you’re like, yes, this has to run fast. So we know what the machine wants to see.

So the question is now: Can the compiler be smart enough to translate between these two levels of abstraction, what the user wants and what the computer wants? And that in a nutshell, is everything that Julia does, and you could not foresee the future.

So we said you’re going to build this thing for ourselves, which means it has to run really fast on our laptops. Jeff used to use Linux and Stefan and I used to both use Macs so Julia ran on Linux and Macs, on Intel processors for several years. In fact, Windows port did not exist for several years, because there was no one who cared about Windows back then.

And we had a very interesting contribution. When in 2012, we announced Julia to the world suddenly there was an outpouring of interest. So we announced, we wrote a blog post and posted on Reddit and Hacker News. And it just took off for reasons that are not clear to us even today.

But out of that came one of one of my co-founders today in Julia computing, Keno Fischer, he saw that blog post, he was in high school back then. And he used to have a Windows computer. And he said, You know what, I’m just gonna port this thing to Windows, so I can use it. And so that’s how we got our windows port.

SHRIPATI ACHARYA 13:20

So what are some of the very first applications of Julia which people wrote? There must have been a need for it, right? And what did you start using it for?

VIRAL B SHAH 13:33

So the real thing is that Julia is usually about a 1000 times faster than say, MATLAB or Python or R. Because those are 100 to 1000, depending on what you’re doing. So the way Python and R and MATLAB are structured is that you try to spend most of your time in library functions, which are implemented in C so that because the languages are so slow, and they are interpreted, I mean, MATLAB has a JIT compiler.

But it’s not open source so we don’t know what happens under the hood. But Python and R, we know what happens under the hood, and they’re interpreted languages. So they try to sort of make the user spend most of their time in libraries written in C and Fortran, so that the speed does not affect them. But in reality, the speed affects you a lot when you write general purpose applications.

And the first use case is again, like we started this thing, so that it would be fun for us. And scratch the itch that we had. And we were back then in a very academic and research oriented environment. And so the kinds of things we wanted to enable were fellow scientists, people who wanted to write at the level of algorithms but needed higher performance.

And so a lot of the initial people were researchers at universities who became contributors, a lot of them were professors, tenured professors at some of the top universities, which is why Julia is of the quality that it is.

And then once it became a little bit more stable and stability is very relative here, but it was in 2018 that we released Julia 1.0 when the language became stable, but already in 2014, and 2015, people started adopting it in the financial industry. So we started seeing quants writing some algorithms like portfolio optimization, derivative pricing, like small sort of snippets of code, mainly to test the waters.

And then some of them reached out to us saying, Hey, we want to use this at work. But it can’t be an MIT project. So we had to form a company just to sort of support some of our first users. But I would say that even today, we carry out a survey of the Julia ecosystem, asking what’s your background, and today it’s 60%, professionals and 40%. Academics.

SHRIPATI ACHARYA 15:57

If you look at it, are there any particular domains which dominate the use cases, when you mentioned finance, obviously,

VIRAL B SHAH 16:06

Finance is a big one of the bigger ones, pharmaceuticals, in terms of percentage of users, it’s small, but in terms of the interest in the pharmaceutical community, Julia has a lot of interest, for example. And so that’s one of the things we focus on a lot, a lot of data scientists, which is a very broad classification.

But data scientists across the board that use Julia, we see, we pretty much see it in every engineering, every Science, Science oriented, every data science thing. So if you look at our community survey that we do, we just published one at JuliaCon a couple of weeks ago, we just sort of see people across the board, there is no industry where there is no interested engineer. And the ratios of interest are actually the sizes of those communities. There are just a lot of data scientists and a lot of general purpose engineers. So those are the two big ones.

SHRIPATI ACHARYA 17:02

So many of our listeners here would be engineering and tech folks, or starting companies or have started companies or are entrepreneurs themselves. They are obviously looking at AI and ML as a key foundation for many of their products and solutions. So if you’re an AI and ML engineer and want to get started at Julia, what are the first things which I would want to do?

VIRAL B SHAH 17:26

If you are an engineer who wants to start using Julia, chances are you already are familiar with something else, maybe MATLAB or Python or something. So Julia is very similar to both of these languages in terms of its ease of use. So I would guarantee that anyone who is used to one of these languages or even R can probably start writing meaningfully useful Julia code within a week.

And the best place to get started is Google Julia in the browser, or Julia language, usually, I mean, our Google rating is good enough that if you just type Julia, you get the Julia language website on the first day. And then there is training, there are the tutorials, and there’s a learning page there.

So there is a learning page there, which has links to training videos, which has links to tutorials. There is a Juliaacademy.com that we maintain as a community, which is just a sort of collection of very high quality learning materials. So those are all the kinds of things.

One specific thing that I would like to point out that your listeners might actually like, especially those students, is that there is a class at MIT called computational thinking. So it’s just computational thinking.mit.edu. And that class teaches you how to think in computational terms using Julia. And it’s probably one of the best and most fun introductions to Julia. But it’s not teaching you just the programming language, it’s teaching you how to think of solving real problems computationally, and happens to teach it with Julia. So you can have a lot of fun.

SHRIPATI ACHARYA 19:10

Oh, great. I think the thing which caught my attention at the earlier part, where you made a comment earlier, which was like 100 to 1000 times faster, and that might have a lot of implications in terms of the speed of rolling out of features as well as maybe even cloud costs and things like that, which startups are very sensitive to.

So I think it is something which I think every engineer should take a serious look at. So Julia itself as a project was started when you were here in India, how has the participation been from the Indian software community contributing to the Julia project? It’s an open source project and to the community at large.

VIRAL B SHAH19:50

So that’s a very interesting question because often when we think about contributions to open source from India, India does not line up. In fact one of my good friends who I shall not name here, has used the phrase which has stuck in my mind that India is a country of downloaders. And that’s so true.

And so I spent the time that I was based out of Bangalore, I spent a lot of time, going to all the open source conferences and talking about Julia and talking to people about open source contributions, especially if you’re a student, or if you’re someone who has some time on their hands, wants to pick up a new project, in terms of sort of hand holding them in terms of how they could contribute.

And we formed a pretty sizable community in India. Largely I would say because of the efforts that I put in back then, and then the community that formed around me and what ended up happening was that the word got around to all the engineering colleges, that Julia is sort of an interesting project that people can contribute to, And by the way, lots of Indian students and Indian software community is contributing to it. So that paired up with the fact that Julia got accepted as a Google Summer of Code project.

And every year when we do Google Summer of Code, more than half the participants are selected. We get a huge number of people admitting, but over half of them are from Indian universities from across the world. And these are chosen based on merit. So I feel like at least in the case of Julia, it has been very special in terms of having a much higher rate of contributions from India, then I think many other open source projects might have, but I have no way to justify this or put real analytics on it. But just anecdotally.

SHRIPATI ACHARYA 22:00

Well, that’s really good to hear. I think it’s more about general awareness of these things. And, of course, it’s not for lack of talent in the country India has got a great pool of software talent, a very deep pool. And I think getting involved in these kinds of things is a feat and I truly believe is only going to increase that statistic that they are a very key part of a vibrant Julia community.

So let me switch gears a little bit. And tell us about your company, Julia computing. For what I understand, it started in 2015. So what was the motivation behind that? And what made you say, this is an open source project, but I really need to start a company now.

VIRAL B SHAH 22:42

Yeah, so the company was sort of a response, like I was telling you, the company was a response to market demand. So we were always profitable from day one, because we had people who wanted our services. So we incorporated the company back in 2015. And started out with our first few customers, mostly in finance back then.

And what we realised was that the traditional thought process around business models in open source was, you sell enterprise support and consulting. And both of those are very human intensive, the service industry, as well as supporting customers.

And, of course, you have to do that when you’re starting a sort of a new and young programming language, and everything is changing all the time, and there isn’t enough talent in the market. But what we realised is that it took away from our time to contribute to Julia, because every time me or my co-founders were writing code for customers, we were not writing code for Julia itself.

And many of those engagements were very enlightening in terms of understanding what the customer demand was, but we realise that we do need to build a product oriented company. And now that’s easier said than done. But it at least became very clear to us in the first year of operation.

And Shripati back then, I spent a lot of time also talking to you, as well. And Sanjay Swamy was someone who I knew from the Aadhar days, so I kind of understood what it took to start a company and what it took to make it successful just from those early conversations. So finally, in 2016, we decided to raise our first round of capital. And we did that with General Catalyst and the Founder Collective, of course, we were still thinking that, okay, it’s open source.

So we’re just going to follow the tried and tested path of what Red Hat has done, what MongoDB has done, what elastic has done. And if you look at all these companies today, they’ve taken so many different paths. And people like to call it the open source business model. But as you get close to each of the companies, you’ll find that actually there is nothing common around their business models.

And then every open source company has had to figure out what its path to market was. And increasingly, a lot of them are focusing on the cloud as the way to build a large scalable business. But the challenges being faced by elastic, especially with AWS, and the incentives and the structures are sort of headline news nowadays, at least in tech media. We took our time. So we raised our first seed round in 2016. And then we subsequently raised series A in 2021, which is uncommon for a startup.

But it was not because we were growing too slowly or we were sort of not knowing what to do, we really wanted to be sure about what that business model is. And because we were very frugal in terms of how we set up our business and how we ran our business in the early days, we were able to give ourselves the time and also able to grow by finding other sources of funds that were non dilutive customer contracts, r&d grants, and so on, and so forth.

So we were able to grow the company to what would have been a traditional series A level company with early product market fit. Team of about 40 people had like 10 or 15 big enterprise logos, all of that stuff. But the business model was not yet clear to us. For the first few years of the business, we also spent a lot of time taking Julia to 1.0, which was only in 2018.

And the focus on business and customers, even though as we were building it, took us some time to figure it out. But now, it’s very clear to us what our path is, and why this can be a breakout company in the technical computing arena. So just for your readers, for your listeners, I’d like to point out that the total addressable market for technical computing software today is about $10 billion in annual recurring revenue.

And there are all the languages that we talked about, some of which are commercial, but there are companies that do engineering simulation, pharmaceutical simulation, chemical simulation, and multi physics simulation, all of these put together, computer aided design, it’s a very large circuit simulation. It’s a very large total addressable market.

And what we realised is that, while Julia gives all of our users superpowers, like the 1000x that you also brought up, we asked the question, What if we could also leverage that same 1000x that we give to our customers ourselves? And what can we build on top of Julia that could actually be something that could be game changing.

And here’s what we came up with, so we said that we are going to build Julia hub, which is our cloud platform on which we are going to provide the best Julia experience anyone can get anywhere in the world, you get there, you fire up VS code, you write all your Julia packages, everything’s available, you click a button deploy to GPUs, or 1000s of CPUs, scale it up. So basically it gives you a supercomputer at your fingertips to someone who’s not DevOps, familiar with DevOps, someone who’s an engineer or someone who’s a data scientist.

And then we took it one step further, and we said, we are going to build domain specific software on top of it, whether it’s in pharmaceutical simulation, or engineering or circuit simulation. And so we took that path of building a suite of Julia powered, domain specific , industry specific applications on the cloud as a path to market. And that is how we define ourselves as a company. So there is a huge change in that five year period, in the six years since we founded the company.

SHRIPATI ACHARYA 28:59

Got it. So if I may in a paraphrase, just say there are two aspects of what Julia computing offers. One is it offers this entire environment, which is running on the cloud, on which you as a user could actually build something, and run the thing that you don’t have to host it yourself. It’s already hosted, and everything else and then you pay.

The other one is actually using the applications which are built on top of this Julia cloud, which also have been built, some of which have been built by you, and then use those applications directly. Is that the right way for me to think about it?

VIRAL B SHAH 29:40

Absolutely. So one is the platform and the other is the applications on the platform. And we decided to build some of these. The first one that we built in pharmaceuticals is in partnership with a company called Pumas AI, and the software is also called the same name as Pumas AI.

But then, as we sort of started building the first one, we realise that we actually have the capability to build others as well. And so the second one we started building is called Julia Sim, which is an engineering multi physics simulation system. And the third one that we started is in circuit simulation, it’s called Julia spice, which is just for electrical engineers.

So we sort of put these things together. And we imagine that this will actually potentially become a marketplace in the future. However good the Julia computing team is it will not have the world’s best experts at everything.

Not even MIT has the world’s best experts, nor does Google. And so really, in order to bring the superpowers that Julia provides to the world at large, this has to be a platform and a marketplace. And that’s sort of the direction we’re heading in.

SHRIPATI ACHARYA 30:49

Got it. So if I’m a Julia user down the road, just like I go to an AWS marketplace, or a Salesforce marketplace, or what have you, I’ll go here and see what are the applications which I need and keep that together. So what are some of the applications that entrepreneurs are creating today? From a startup perspective, which you’re seeing and what are some of the emerging areas?

VIRAL B SHAH 31:17

Yeah, that’s a very good question. So because, like I always like to say, Julia gives the programmer superpowers, and those superpowers lead to disruption of old markets or creating new markets, because the world is always evolving.

And while Julia is used in over 10,000 companies, by my count, I know of at least half a dozen startups that only were possible because of Julia. And their entire stack is built in Julia today. And people have built all sorts of things, there are healthcare applications, pharmaceutical applications, machine learning applications, energy analytics, and so on and so forth.

There are just so many different things that startups today are building in Julia, that is just amazing. I mean, I never just imagined. For me, going from an open source project to starting Julia computing was such a lift. But I was just thrilled to see that not just Julia computing, but there are so many other companies that are being made possible, because of Julia.

And in fact, this is something that I have not shared with most people yet, but since you brought it up, I’d love to point out that there’s actually several hundreds of millions of dollars that are going into Julia, or into startups that are built on top of Julia today.

So it’s, if you’re an entrepreneur, it’s a great way to get ahead of the game, cut your costs, save money on your cloud, get the best engineers and actually build something that is game changing in a very short amount of time.

SHRIPATI ACHARYA 33:02

I mean, that is really exciting. So we are running out of time. So I’m going to close a two part question. One is, what is next for Julia in terms of the direction? And of course, what is next for Julia computing the company?

VIRAL B SHAH 33:17

Yes, so Julia is an open source project. And so we do not publish a big roadmap, it’s based on what contributors bring to the table. And the contributors can be companies, they’re individuals, but they can also be universities. So Julia project, the one thing I can say is that we have big plans for what we think Julia should be able to do.

One of the major things that we just are bringing into Julia is this functionality called automatic differentiation. Everyone has read about the famous blog post on software is eating the world and actually that one and sort of the other one on software 2.0 which is, Andrej Karpathy’s blog post on how AI and machine learning are reshaping software engineering, where you might sort of use data and models rather than write programmes by hand. And Julia is sort of at the forefront of enabling that whole concept.

Because of this idea of automatic differentiation, what Julia users can do is they can write their Julia code and then automatically differentiate it just like, take a derivative.

And for those of you who might remember some of your high school calculus or college calculus, if you can take derivatives you can optimise it, if you can optimise you can build a portfolio management tool, you can build a neural network and detect images or you can build a Tesla’s self driving stack, or if you’re a scientist, you can optimise the dosage of a drug or you can design the perfect chemical reactions for your new batteries and so on and so forth. You can model climate change, and figure out how that’s going to impact everyone.

So we mapped all these key scientific capabilities back to what language level innovations, we always ask the question what language level innovations will be required to power the next generation of scientific discovery. And that’s where the Julia language is headed. And automatic differentiation is the current big project in that space. Many other projects around reducing compilation time, reducing compilation delays, all that sort of stuff, which is largely engineering oriented.

But for Julia computing, again, I map the same vision, but it’s a business lens. There are so many difficult problems in the world that we need to address. And there are so many of them. At the same time, climate change, we’ve got a pandemic on our hands. I mean, we need to move to renewables and they are all interlinked.

I mean, there are water scarcity problems, there’s floods, there’s all this climate mitigation stuff that is coming up. And we believe that Julia has a role to play Julia computing has a role to play. And by building a suite of domain specific industry specific solutions, and by making simulation easy and accessible to all of our business customers, we think we can bring them tremendous economies of scale. And that is the focus for business for Julia computing.

So if you want to check it out, Juliahub.com is a cloud hosted platform, you can go there and try out some of these domain specific tools, and simulation tools that I mentioned. But if not just try out a Hello World and try out a simple machine learning application and the first $25 are on us. So no cost for trying it out.

SHRIPATI ACHARYA 36:46

Wonderful. I think that given how core high performance data crunching is to the principles on which Julia the language has been built. It seems to me that any engineer, and every entrepreneur, honestly, who’s looking at large data sets is looking at creating models on top of them must actually check out Julia and then the application that is available on Julia hub. So let me ask you, do you still write Julia code in your spare time?

VIRAL B SHAH 37:19

Absolutely. I mean, that can never get away from that. So we’re just going to ship Julia 1.7 in the next few weeks, or maybe a couple of months. And it is going to have a completely new system for how we do our linear algebra under the hood. And I was one of the major contributors to that. So I actually have a contribution in the next few weeks.

SHRIPATI ACHARYA 37:43

Well wonderful Viral, it’s always a pleasure talking to you. And all the best. Of course, we expect great things out of Julia going ahead and all the best for Julia computing.

VIRAL B SHAH 37:53

Thank you Shripati. Thanks for having me on your podcast.

 

Enjoyed the podcast? Please consider leaving a review on Apple Podcasts and subscribe wherever you are listening to this.

Follow Prime Venture Partners:

Twitter: https://twitter.com/Primevp_in

LinkedIn: https://www.linkedin.com/company/primevp/

This podcast is for you. Do let us know what you like about the podcast, what you don’t like, the guests you’d like to have on the podcast and the topics you’d like us to cover in future episodes.

Please share your feedback here: https://primevp.in/podcastfeedback