I'm the Technology Editor at Forbes India and I love writing about all things tech. Explaining the big picture, where tech meets business and society, is what drives me. I don't get to do that every day, but I live for those well-crafted stories, written simply, sans jargon.
Deepak Vinchhi, Co-Founder & Chief Operating Officer at JuliaHub
A new programming language takes hold probably once a decade or so. And in the world of scientific machine learning, Julia, an open source language, is gaining traction. The language’s co-creators also started a company, JuliaHub, to offer enterprise-grade products.
The company, with centres and team members around the globe, makes complex scientific and mathematical calculations easier for customers in aerospace, defence and pharmaceuticals. Vinchhi explains how, in a recent interview, and spoke about their plans. Edited excerpts:
Q. Tell us about Julia computing language and JuliaHub So Julia, which is an open source language project, was started in 2009 by Viral Shah, Alan Edelman, Jeff Bezanson and Stefan Karpinski. The goal was to address the "two-language problem" in mathematical computing. Traditional languages like MATLAB, Python, or R were not fast enough for complex computations, so users had to rewrite code in faster languages like C or C++. Julia aimed to be as easy and productive as the high-level languages while performing as fast as C.
Today, Julia has become mainstream with over a million users and more than 50 million downloads. There are over 1,000 active contributors worldwide. The language remains open-source and self-sustaining.
To support commercial-grade usage, we founded JuliaHub (formerly Julia Computing) in 2015 as a platform for easy development, computation, and deployment on the cloud. Enterprise customers benefit from governance features, project management, traceability, and reproducibility. We have successfully served industries including pharma, aerospace, and engineering, helping them meet regulatory requirements and solve complex mathematical problems.
On the language front, Version 1.9 was released recently, and at JuliaHub, you announced a new funding round – give us an update.
Julia version 1.9 is a significant release with improved features, enhanced robustness, and faster pre-compiled time. As for JuliaHub, the company, we recently announced securing strategic funding from Aero Equity Industrial Partners and HorizonX Fund in partnership with Boeing. This collaboration aims to strengthen our presence in the aerospace and defence industries.
AEI's expertise in this vertical will be invaluable as more companies seek our help in using Julia's language and scientific machine learning for engineering challenges. Many automotive, aerospace, and defence companies, particularly in the US and Europe, use Julia to solve complex problems and are turning to us for support, including our JuliaSim product.
This investment of $13 million marks Series-A1, an addition to our previous Series-A round of $25 million led by Dorilton Ventures two years ago. In total, we have raised around $43 million over the last seven years, including a seed funding of $4.6 million. The future looks promising as we continue to grow and expand our reach in these industries.
At JuliaHub, we are a diverse team of around 100 people spread across the globe. Approximately one-third of our team is in the US, another third in India, and the rest in various countries across Europe, Japan, and Australia.
Q. Tell us about the markets where you’ve seen significant adoption, and some of your well-known customers Julia and scientific machine learning have seen significant adoption in the US and Western Europe, although usage extends worldwide, including China and India. Our commercial focus centres on the US and Europe, with major pharma and engineering companies, along with top Fortune 500 firms, using JuliaHub, JuliaSim, and Pumas products.
Some notable customers include Pfizer and Moderna, who used Pumas for the infant Covid vaccine, achieving impressive results. Additionally, government organisations like NASA and the Federal Aviation Administration use Julia and scientific machine learning in their research labs. Our customer base continues to grow, presenting exciting opportunities for the future. Q. Give us a sense of how your products help your customers Our focus is on making our customers' work much faster, resulting in significant productivity and cost gains. With Julia and scientific machine learning, we've seen drug development cycles compress by 80 percent in some cases and product development cycles shrink from five years to two years.
We're innovating in the field of scientific machine learning, combining science and machine learning to solve scientific problems efficiently without the need for massive amounts of data. This approach sets us apart in the industry, enabling thousands of times faster and more efficient results.
At the core of our offerings is the JuliaHub platform, facilitating development, deployment, and scaling of Julia programs. It comes with enterprise-grade features for governance, project management, traceability, and reproducibility. Customers can collaborate seamlessly using Git repositories and scale programs to hundreds of cloud cores with ease.
We have three domain-specific products: Pumas, a mature drug modelling and simulation tool for pharmacometrics; JuliaSim, focused on scientific machine learning to solve engineering challenges in various industries; and Cedar, an upcoming circuit simulation tool. Although each domain has unique applications, the underlying mathematics and scientific machine learning principles remain consistent, enabling cross-domain learning and innovation.
Our domain strategy allows us to better understand customer problems, providing more value compared to pure Julia usage. We're on an exciting journey, with each product maturing and making its impact in the market.
Q. Give us a sense of your business model. Is it like a SaaS company? Our entire JuliaHub platform and these domain products that sit on top of JuliaHub are offered on the cloud. So it's a SaaS model, self-service model. But when it comes to enterprises, they typically also want certain other features that we also provide.
We can also set up our products on-premises, on their own cloud instance. Or it can be in their data centre.
And for defense and some of the government requirements, they also want air-gapped solutions where they want to install the whole JuliaHub and JuliaSim in a completely secure environment. So we have an offering called JuliaHub Air where it can be deployed in a secure environment.
Today many well-known system integrators, including the Indian IT services companies, are familiar with Julia and have teams working on it.
Q. And over the next 12-18 months, what might be your big priorities? Our top priority would be to make JuliaHub and JuliaSim successful with large customers. Now there will be a lot of focus on aerospace, defence and pharma. So these three, we already have some customers, and if their adoption of our product grows, they we will grow fast as well, because these are very large companies, and the opportunity is huge.