Why UBC student Veer Bhatia is bullish on Julius AI becoming the world's ‘most useful AI platform’

The shift to AI platforms is ending the era of shallow tech as deep tech innovations take centre stage.

One such platform paving the way is Julius AI, an AI data analyst tool that makes it simple for anyone to analyze large amounts of data, create charts and graphs, gather insights, solve math problems, and generate reports within seconds — regardless of one’s technical expertise. What sets Julius apart from similar tools beyond its UX and speed is that it offers more accuracy, built-in privacy, and no rate limits at lower costs.

According to a new PwC report, sectors more exposed to AI are experiencing almost five times higher growth in labour productivity, yet the rate of adoption in Canada remains lower than in the U.S. and U.K. Julius is a promising tool that can help change that as it’s quickly building momentum in many ways.

Launched in 2023, Julius has over 600K active users, including data and sales analysts, researchers, healthcare providers, environmentalists, employers, and institutions like Harvard Business School. Julius has also partnered with Wolfram Alpha, a search engine for computing answers and providing knowledge, to supercharge their own engine. To date, the startup has secured funding for an undisclosed amount from:

• Y Combinator, the famed startup accelerator and VC firm;

• AI Grant, another accelerator by former Y Combinator partner, Daniel Gross, and former GitHub CEO, Nat Friedman;

• Founders and CEOs from the likes of major tech companies including Palantir, Notion, and Zapier as well as;

• Researchers at OpenAI, the AI research company behind ChatGPT, and Anthropic, an AI safety and research company.

To scale Julius, founding engineer Veer Bhatia, rejected offers to join the big-name tech companies and dropped out of UBC’s computer science program to move to San Francisco. There, Bhatia is now working on the platform full-time alongside the rest of his team. Ahead of the move, he and I chatted about his journey of:

• Learning to code from YouTube before working as an ML and AI intern at Amazon;

• Meeting fellow founder, Rahul Sonwalkar, by chance on Twitter;

• How he and his team quickly acquire hundreds of thousands of users;

• What can be expected next from Julius as it’s poised to make history and;

• How his Vancouver roots helped him get to where he is now.

Let’s start with your early life. Growing up, what were your goals and aspirations?

I grew up in Squamish, B.C., with my mom and grandma, who had immigrated from India just before I was born. 

Growing up in a single-parent household, the one thing that stuck with me was just how much my mom had to work in those early years — often working multiple retail jobs a day to make ends meet. 

While my mom was at work, my grandma would look after me, telling stories of her life back home and instilling in me the values of hard work, perseverance, and the importance of education. 

They sacrificed a lot for me and made sure I understood the importance of working hard, so it's always been a huge goal of mine to make them proud and, in some way, be able to pay it back for their love and support. 

My aspirations growing up oscillated between the intersection of the arts and the sciences. At different points growing up, if you asked me about my goals, it would’ve been things like poet, physicist, philosophy professor or mathematician. The closest I came to was the latter: I ended up doing a degree in honours mathematics at UBC. 

The thing that always guided me, even when it wasn’t clear what I wanted to do, was what I learned from my mom and grandma: to always learn as much as possible and use that knowledge to help others.

Who or what had the biggest impact in shaping the direction of your career?

I think at some point, I went looking for a role model or a father figure in books. 

By the time I got to high school I had read or was reading pretty much every biography I could get my hands on. My favourites were biographies on Walt Disney, Steve Jobs, Henry Ford, Larry Ellison, and Phil Knight. 

Around that time, I also read a book called Zero to One by Peter Thiel, which is all about innovation, specifically startups. I think some combination of these events made it all click for me in terms of what I wanted to do. 

I realized that I needed to work on important problems that would help real people in quotidian life, just like the people in the biographies I read about, and that the best way to do that was through a startup. 

I think we all want to contribute something to the world, not just as a way to express ourselves but to show appreciation for all the contributions that came before us. I couldn’t write or sing, but I did manage to teach myself to code by watching YouTube tutorials during lunch at the library and after school, so I decided to go about it by creating software. 

You studied and did internships in different cities. What made you decide to come back to Vancouver?

As a tech hub, it’s a great place to be. 

There's a large presence of world-class universities, some of the biggest tech companies like Microsoft and Amazon, and amazing up-and-coming startups. 

There’s also always a founder meetup or AI hackathon to attend — no matter the time of year. 

But the biggest factor that made me return to Vancouver was the people. Some of the kindest and friendliest people in the world are here. 

I also think there’s a real calming vibe to being in Vancouver. I find it really easy to focus here and work. Maybe it’s the beautiful scenery we’re surrounded by or the rainy weather that makes us all stay indoors most of the time. 

Either way, I miss it every time I leave, and I’m sure I will miss it more in the future. 

What was most exciting about your internships that was especially valuable to building Julius?

I think the most exciting thing my internships taught me was what it’s like to start from scratch and release a product into production. 

At Amazon, the code from my project is still being used today by millions of people, but with that came many hurdles. Big companies have strict guidelines on software quality, and rightfully so, given how many people rely on them to work each day. 

Building around these criteria forces you to approach problems with the customer at the forefront and to think deeply about the design of each aspect of your application, its reliability, and its impact.

When and how did you meet Rahul and embark on the idea of Julius?

I actually met Rahul through Twitter during the pandemic. 

We were both at home, me in Canada and him in the States. We hit it off right away since we worked on similar problems in ML. Twitter is not what it once was, but it was a great place to talk about tech and AI with other like-minded folks at the time. 

A couple years later, Rahul started working on an open source project to convert natural language to SQL. I thought it was super interesting and began contributing to the project while studying at UBC. 

From there, we started collaborating on a series of projects centred around data and AI that eventually turned into what we now call Julius. 

Walk me through the journey of starting Julius to acquiring your first users.

In early 2023, we initially worked on an open-source project called CensusGPT that allowed users to ask natural language questions such as “What’s the most diverse neighbourhood in San Francisco?” 

We used OpenAI’s GPT-3.5 API to convert these queries to SQL code that would run over a census dataset and return the answer to the user with some cool UI, like a map highlighting the pertinent area. 

This project led us to HoopsGPT, which focused on a similar idea but on NBA data. These sites collectively attracted over 200,000 unique visitors, but we knew they needed more resources and a sustainable business model. 

Eventually, we launched a GPT store plugin called “Chat with your data” that we saw a lot of traction with. Users were saying we saved them hours of time analyzing spreadsheet data and that’s what led to the creation of Julius AI. 

In terms of our tech stack, we’re a scrappy team that likes to move really fast, so it’s as simple as it gets. We use Next.js and Python, and that’s pretty much it. It’s nothing fancy, but shipping products is more important than over-engineering.

We acquired our first users using the GPT plugin store. Though it no longer exists, it was a great way to onboard users with zero distribution cost and introduce them to the magic of Julius.

Now, we’re focusing heavily on getting Julius in the hands of influencers around social media such as YouTube and Instagram as a way to grow and share their reviews of the product with their respective audiences. 

What has been most rewarding about building Julius?

It’s got to be the customers. 

Every day, I talk to people from all over the world who are working on real problems. It’s extremely rewarding to learn from them and hear how we're able to help empower them.

For instance, I had a Zoom call with a doctor from South Africa last week who was using Julius to analyze data from her laboratory to research mother-infant cardiac synchrony. The visualizations she had managed to create were incredible, and they were all ones we’d never envisioned. She also mentioned that even though the data analyst on her team had left, she could derive insights using Julius. 

It’s not just individuals, either. This semester, Julius was used at Harvard Business School, Rice University, Brevard College, and other institutions to teach entire courses on data analysis, all powered by AI. 

What were the biggest challenges in building Julius? How did you overcome them?

The biggest challenge is always growth and distribution. 

You can build the best product ever but it doesn’t matter if no one is using it. Getting the word out there to the people who need it the most is a huge priority.

The best way we’ve overcome it is just by talking to and learning from our users. 

What are the most common use cases you’re seeing on desktop and mobile?

The most common use cases we see are data analysis, solving problems in math, chemistry, physics, statistics, and programming, and chatting with documents. 

One big thing on mobile is how much users love taking pictures of problems — from data to math — and getting Julius to help walk them through how to solve them. 

Recently, students in our support channel shared how they used Julius to help them study for their statistics finals to supercharge their learning like a tutor and that they managed to score A+s across the board.

What are the newer and more unique use cases that some may find surprising?  

We're seeing some really unique use cases, such as people using Julius for video editing, video compression, and file editing. 

Because Julius is essentially a code monkey in your pocket, you can tell it in simple words what transformations you want to make to your files, and it can execute code to do exactly that.

How does Julius outperform ChatGPT and other AI tools out there?

Many of our users use Julius in conjunction or parallel with ChatGPT. We’re huge supporters of all the newest AI products. 

Currently, we’re the only platform with a powerful code interpreter that powers both OpenAI GPT-4 Turbo and Claude Opus models. Our code interpreter is hyper-optimized for analysis and problem-solving. We’re always improving our tool selection to keep it the most advanced and the most affordable in that sense. 

On ChatGPT, you need to pay $20 to access their latest model, GPT-4,. On Claude.ai, you also need to pay $20 to access their latest model, Claude Opus. These platforms have rate limits that prevent users from sending too many questions in the span of a few hours. 

Julius allows users to chat with both models — 15 messages for free users and unlimited for paying ones — with no rate limits. 

In addition, Julius offers an advanced code interpreter that enables advanced data analysis, seamless visualizations, and problem-solving for even the toughest math problems. It's sort of like the best of both worlds, where you can chat with these amazing models and run code, analyze, and chat with files.

There’s always been a lot of competition in technology, but we plan to stay ahead by relentlessly innovating and shipping things really fast. I always bring up two quotes: “The quick shall inherit the earth” and “If we do everything, we win.”

How does Julius plan to remain competitive in the future?

A lot of it is figuring out what our users want before they do and innovating precisely on that front. We’re constantly shipping new things every week. 

For example, last week, we announced R with Julius. This highly requested feature lets you execute R programming language natively inside the platform. R with Julius is the first of its kind for an AI code interpreter. 

Typically, AI code interpreter tasks are performed using Python, which isn’t always the language researchers and scientists working with data are familiar with. R is one of the world's most powerful analysis languages, and millions of people struggle to learn to use it every year. We hope to completely change how R is used in classrooms and labs to visualize data. 

As a startup, we cannot rest. We need to constantly work towards building a better product that excites our users.  It won't be easy, but as long as we obsess over our users' needs and make them happy, we think we’ll end up in a good position.

How has Julius approached improving the product while marketing it to reach more users? 

We like to think growth and innovation go hand-in-hand. 

As we gain more users, we learn more from them, and using that feedback, we're able to keep improving the product, pushing past the current frontier. 

In fact, much of our growth and marketing impact has come from the improvements we make to the product. 

What was the moment that made you decide to drop out of UBC and pursue Julius full-time in SF?

At some point, I realized I was spending more time on Julius than on schoolwork. 

It just became clear that while we were on this exponential trajectory, there was still a lot to build, and it was important for me to prioritize that and make a firm decision to go full-time.

I think it was more of a gradual decision than a sudden one. I saw how many users we were able to help over the past few months, and it didn’t make sense for me to work on anything else.

What were your family and friends' reactions? Did you face any criticism? If so, how did you deal with it?

My friends and family have been super supportive. I think the only criticism they give me is not to work as hard as I do. I’m basically working intensely on Julius 24/7, so it’s hard to always make time for everyone. 

A big thing I’ve been trying now is to allocate dedicated time away from the screen. Apart from exercising, I'll hang out with friends and family. While a perfect balance might be hard, it’s still important to spend quality time with people in your life as much as possible while working on complex problems.

I think the toughest part for me personally was having to turn down offers to join the big-name tech companies in Seattle and Silicon Valley to keep working at Julius instead. When I tell people the companies whose offers I rejected, they all look at me sort of crazy at first until I walk them through why I wanted to work at Julius instead. With how tech hiring is right now, it’s something that I had to think hard about. 

What led me to overcome any doubts is that I strongly believe that now is the greatest time in history to build a startup. The recent innovations in AI are a watershed moment in history. Not only is the potential for this tech bigger than the internet, but I think it’s also the last time in history any human will be able to have an outsized impact. 

With the current rate of progress, it’s increasingly inevitable that the world changes, that most of our interactions come to happen with AI, and that we’re able to automate away a lot of the things we’ve grown accustomed to doing ourselves — like building startups.

What’s the Julius team and office setup like in SF?

We have five people on the team, including myself. The team is small but full of the smartest and hardest-working people I know. I learn a ton from each of them every day. 

Our office is in San Francisco’s Hayes Valley neighbourhood which has become a bit famous now because of how many AI startups are currently residing within it. 

Did you ever think that you’d one day become an entrepreneur?

I always knew I wanted to build and lead my own company at some point. I think that’s always been something I wanted to do: build a company that would create great products and be able to help many people for generations.

I think in recent years, the term ‘entrepreneur’ has sort of been hijacked by people who want to launch a startup and then sell or go public to cash in and move on. That’s never been a goal for me. 

I would much rather contribute to the flow of our world and add to the legacy of those who went before me by building a company like Apple or Intel. Both not only exemplify the Silicon Valley entrepreneurial spirit but also stand and last for generations as a beacon of inspiration.

What have you enjoyed the most about your startup journey so far?

I really enjoy how much ownership I have of the projects I work on. I’m able to make a lot of large decisions about parts of the product, which is super important to me. 

Another aspect I enjoy is how quickly I can move on things. Facebook’s old motto was “Move fast and break things.” I think that embodies a lot of our mission as well. Our advantage as a startup is that there’s no bureaucracy slowing us down. All we have to do is write code and prove its worth in a production-level environment. 

While this is intrinsically challenging — having to work on hard problems and put solutions out there in the real world for testing — it's also incredibly rewarding to ship something really great. 

What limiting beliefs, if any, did you have to overcome?

I think one of the biggest pitfalls you naturally fall into when starting an early-stage company is comparing yourself to others who've been doing it longer than you. That’s something I’ve had to overcome and it’s a lot like self-belief in a way. 

Instead of comparing ourselves to other startups, we try to compare ourselves to our previous selves and how much progress we’ve been able to make in the last period.

What’s been keeping you grounded and focused — a book, podcast, or hobby?

I love reading the essays by the founder of Y Combinator, Paul Graham. 

Though he writes broadly on many topics, his advice on startups has always been particularly helpful in keeping me grounded. 

I still read a lot of biographies, too. I just finished the Walter Isaacson one on Leonardo Da Vinci and am about to dive into one about Alexander the Great.

What are you most proud of being part of founding Julius so far?

I’m most proud of the progress and impact we’ve made in under a year. This time last year, we hadn’t written a single line of code for Julius. Today, the product helps hundreds of thousands of people each day and saves them countless hours. 

Right now, students are telling us daily how they can’t wait to subscribe again once the school year starts up again. Researchers and academics are also subscribing to power their summer research endeavours. 

I'm grateful for the quality and amount of users we’ve been able to help through Julius with real-world problems. The craziest part is we’re just getting started. 

What are you bullish on what Julius will become in a year and five years from now?

I love this question, not only because it’s something we think about but especially since we overuse the word ‘bullish’ at work. This is exactly how we’d phrase the question to ourselves, too.

In a year, we want to have our superintelligence that can do math, data analysis, physics and so much more, in the hands of millions of people. Julius is already transforming the way some classes are being taught at a university level, but there’s a huge untapped area of users that could immensely benefit from this technology. Our goal is to get to them. 

Five years from now, I think tech will look a lot different around the world. The interface and how we interact with software is something I believe will especially change. Julius might just be a specialized intelligence that you hop on a video call with each morning, give it a list of things to do —  from programming to data analysis —  and it comes back to you within a few minutes with everything done perfectly. 

Models such as GPT and Claude are getting smarter at an exponential rate. The rapid acceleration will allow Julius to become more intelligent and capable, maintaining it as the most advanced tool selection and, eventually, the most useful AI platform globally. 

What is the biggest impact you envision Julius will create for its diverse users?

I love the wording on this question as well. Our users are super, super diverse. There’s professors in academia, PhD students, college students, freelancers, business owners, and hospitals. 

The biggest impact we envision is empowering our non-technical users to analyze data, solve problems, and generate insights like top analysts in the world are able to. 

Our vision is that eventually everyone has a Julius code monkey in their pocket that they can use to solve problems, analyze data and write code — no matter where they are in the world. 

What can people expect to come out of Julius over the next several months? 

We're hyper-focused on delivering the best computational AI, which informs a large part of our upcoming roadmap. 

We'll be integrating support for many more models from providers in the coming weeks, such as Cohere and Google Gemini. This will allow users to evaluate each independently with our built-in tooling and pick the best one for their use case.

Support for GPUs is also coming up, which will help power even the most advanced research workflows on Julius. It will also allow anyone to do things like train models or perform complex scientific simulations — all without having to write a single line of code.

We’re constantly thinking and rethinking the way users interact with AI tools and shaping our platform around a belief in generative UI and a true AI-native interface. We have a lot of things in store we can’t wait to share, but most importantly, we love hearing from our users about where we can improve all aspects of the product.

What are you most thankful for about your life in Vancouver that helped you get to where you are today?

I was still in high school when I got accepted into the League of Innovators (LOI) batch 8 in Vancouver to work on a startup I had founded by myself. It was called Orwell and focused on mitigating social media addiction and improving community mental health. 

I’m thankful for that experience, as it shaped a lot of what I ended up working on for the next few years. LOI took a chance on me, allowed me to pitch, taught me the ins and outs of startups, and gave me the confidence I needed to pursue a career in tech and work full-time in startups.

What advice do you have for other students in Vancouver who are thinking of dropping out to focus on their startup full-time? 

I think that's a super personal decision, and my advice would vary from person to person. 

I recommend following your natural curiosities and letting that guide you both in school and outside of it.

Is there anything else you want to add or feel is important to share with VTJ readers?

It feels like we’re in a powerful moment in our country. There are many doubters, a rough economic climate, and an ever-looming housing crisis. But there’s also a lot of hope, and it’s up to us to shape the future of Vancouver and Canada as a whole. 

It’s important for us to be ambitious and share our learnings with each other to push forward using tech. You and VTJ write incredible stories each day and allow everyday people like me to be able to share my own in hopes that it’ll inspire others. 

I never went to a great high school, never went to Waterloo, never studied the right major, lived hours away from campus, never got any referrals or shortcuts, nor did I have an easy time securing internships between semesters. I worked really hard to be able to get a chance. 

One thing I always kept with me was an unwavering belief that I could learn anything, and my one ask is that readers do the same. Continue to inspire, and lead, and I hope to read about you here soon too.

Don't miss out on what’s happening in your backyard. Subscribe or become a member of Vancouver Tech Journal.