Learning comes in many different forms. Athletes learn via practice. Musicians by sound. Large language models with data. But interns at Snowflake? They learn through meaningful, hands-on experience — from owning projects that go into production and collaborating with industry titans to mentorship that molds them into the next generation of innovators.

So what do you get when you combine next-gen tech with next-gen talent? None other than Abishek Sridhar. Below, the former intern and current AI & ML software engineer sat down to share lessons learned from his internship, navigating the fast-paced world of AI, and what it’s really like to level up (and snag a full-time engineering role) at Snowflake.

Let’s chat about your intern days — what made you decide to intern at Snowflake?
During our initial conversations, my recruiter thought I might be a good match for the ML team based on my experience. As soon as I met with the hiring manager, I immediately knew that the work they were doing, and that I would be doing by extension, was incredibly interesting. It was a very fast-paced environment, the team was ambitious, and I remember walking away excited about the kind of ownership I’d have over my intern project. A lot of interns at big companies don’t get the opportunity to work on such an impactful thing, so that was a big draw.

Tell us more about the work you did during your internship.
The team was doing a lot of exploration around natural language processing, so my project involved working on obtaining semantically meaningful row embeddings. This was back when there was no Gen AI hype — it was just starting to emerge. My focus was on developing embeddings to retrieve results similar to how Cortex search works now, but more generally. The idea was to create embeddings that could encapsulate an entire row of data — both text, numeric, categorical data, and so on.

Mentorship can make or break an intern experience. When thinking back on the mentorship you received, how did it help shape your skills, perspective, and overall experience?
I had interned before, but this internship felt different — it’s where the importance of mentorship became really clear to me. My mentor, Luke Merrick, was amazing. He was very supportive and laid out exactly what was expected from me for the project and broke everything down. He made it clear that this was a research-oriented project, so I wasn’t expected to come up with production-level code, but whatever I worked on would eventually be translated into production. Having clear expectations meant I always knew where I stood, whether midway or at the end of my internship.

One of the biggest takeaways for me was learning how to write clean, modular, and extensible code. I also learned the importance of reproducibility in industrial experiments, as well as the right kind of documentation. It’s become even more valuable as I continue my work, but I owe it to Luke for giving me so much valuable insight, from the technical side of things to even just his mindset around collaboration. His long-term thinking helped shape my perspective on how I approach my own work and the company’s mission to this day.

You’ve since been converted to a full-time employee (congrats!). What made you want to return as a bonafide Snowflake?
Throughout my internship, I was able to understand and align on Snowflake’s long-term mission in the ML space. But one of the great things my mentor did was keep in touch with me after my internship wrapped and I finished up my degree. He kept me informed about his work, priorities within the team, and even leadership changes, so when I did come back, none of it was a surprise. To be honest, I never really felt like I left Snowflake, which was really important to me.

Can you give us a taste of what you’re working on now?
So right now, I'm part of the Cortex Analyst team. Specifically, I'm working on coming up with models that would be able to generate the right SQL when customers ask questions in natural language. So it's like customers can just ask anything, and then there's this whole ML agent system behind the scenes, doing all the work for you, and then translating that to SQL, which the customers can then immediately run and see the results for themselves. It’s not an easy problem to solve, so the team’s ambition I mentioned is still very much intact.

How do you feel the work you're doing now impacts Snowflake and the industry at large?
Even though text-to-SQL efforts have always been out there — there are multiple models, companies, and startups trying to develop coding agents and increase efficiency — I think Snowflake still has its own niche space. It's inevitable that companies will store data, which can range from customer feedback to data they want to use to train their own models, to whatever log data they come up with. The way AI is structured on top of this data, having access to all this data, and empowering customers by providing easy ways for them to interact with the data or use the data for downstream tasks is really transformational.

Specifically, regarding my team, coding agents have always been out there, but to create this as a viable product for customers, you need to achieve performance that is close to perfection. This is one particular use case in AI that needs a perfect score to be really reliable and trusted by customers. We're moving towards it, and if we get anywhere close, it'll be revolutionary because of the performance level of that magnitude. So, big things are on the horizon, for us, our customers, and the industry in general.

Now that you have the wisdom of retrospect, what advice would you give aspiring Snowflake interns?
I think the first thing is to definitely try to talk to a lot of people, even before you start your internship, or even during the matching phase, to find the right team that you're interested in. That goes a really long way in setting the right expectations and creating a happy mindset. You should really enjoy your work, no matter where you are in your career.

Once you’re matched into the right team and begin the project, I also think it's really important not to be siloed. Interacting with a lot of people was game-changing. You get so many insights about features and other teams, but it’s also about developing interpersonal skills. A lot of software engineers sort of keep their heads down and focus on coding, but an internship is also about enjoying the process, attending social events, and building relationships. And that way when you come back, everyone will already know you, and that’s really special.

We’ve got the next-gen tech, but we’re always looking for next-gen talent. Learn more about University Recruiting at Snowflake and check out open roles here.

Abishek Sridhar is a software engineer on Snowflake’s AI & ML team in Menlo Park, CA. Prior to joining Snowflake as a full-time engineer in 2024, he interned at Snowflake and held graduate assistant roles at Carnegie Mellon University. He received his Bachelor of Technology from the Indian Institute of Technology, Madras in 2022 and his Master of Science in Computer Science from Carnegie Mellon University in 2024.