This as-told-to essay is based on a conversation with Lambert Liu, a software engineer. The following has been edited for length and clarity. Business Insider has verified his employment and academic history.

For most computer science graduates, it’s a no-brainer to work for Big Tech.

Most of my classmates were drawn to Big Tech companies like Google, Meta, and Amazon because they promised prestige, stability, and a structured career path.

But I found myself falling into a second group of college students, one that actively seeks job opportunities at startups for the steep learning opportunities and potential equity upside if the startup goes public.

I reached that decision after doing internships in both Big Tech and startups.

I did two internships at Google during my sophomore and junior years in college.

When I interned at Google for the first time, I really liked it. But when I went back for a second round, I thought my growth there was plateauing. I didn’t see myself working there in the long term.

At the end of my junior year, I did an internship at Replit, an AI software development startup. That experience was refreshing because I got to lead impactful projects. I realized I wanted to work at a startup, and that led me to my first job at Graphite, an AI code review platform.

Here are the top tips I have if you want to land a job at an AI startup.

Big Tech experience helps

If you are like me and want to give startups a shot after interning only at Big Tech, don’t worry. You don’t need past internship experiences at startups to work at one.

Interning at a Big Tech company helps demonstrate to employers that you have a strong overall technical foundation. You will know how to do great technical design and great testing. Your stint with Big Tech tells recruiters that you are capable of writing clean code and shipping reliably.

It’s good to have startup experience because you will be more used to dealing with ambiguity and thinking quickly on your feet. But that can be easily solved by working on your own personal projects, which takes me to my next point.

Build more projects

I personally worked on a lot of passion projects in my downtime when I was not working. Working on those projects not only developed my skills, but it also helped strengthen my approach toward solving problems.

In fact, those projects do not have to be AI-related. You can use AI tools to amplify your productivity as an engineer, but you should not limit yourself to just working on AI projects.

Building AI projects also isn’t a prerequisite to working in an AI startup. They generally look for great engineers, and whether you build projects with AI or not, there are many ways to demonstrate your thinking and technical skills.

LeetCode still matters, but not as much

Solving algorithmic and coding problems on LeetCode, an online learning platform, still matters when you are preparing for technical interviews at startups.

That said, there’s a lot more emphasis on one’s ability to deal with ambiguity and tackle non-technical areas like product thinking. This is especially the case since every engineer can use AI to write code.

Working on your own projects will help you strengthen your problem-solving skills. Having to build something new forces you to develop your perspective and taste for approaching problems, which will help you better handle the interview.

Get good at system design thinking

My job interview at Graphite was the first time I was ever asked about system design. That is not usually asked of new graduates. When it comes to system design, companies assess not only your technical skills but also your approach to problems.

I learned a lot about system design thinking when I took a course on human-computer interaction in college. I learned how to scope problems and then build a technical foundation to solve them. The course also gave me some hands-on experience when I built a project.

Foundational courses like algorithms and data science are important, but going into areas like human-computer interaction will be useful when you start interviewing.

Be a holistic engineer

If you want to excel at a startup, you must strive to be a holistic engineer above all else. You need to work at a fast pace. And on top of that, you have to show that you really care about your users.

You can start doing that now when you are interning. Show your bosses that you really care about your craft and want to make the best possible product.

Take ownership of your work as much as possible. At AI startups like Graphite, we move fast, so we are looking for hires who can cope with that velocity and produce high-quality work.

Do you have a story to share about working at an AI startup? Contact this reporter at ktan@businessinsider.com.

For Nadia, Sage Copilot is an essential ally when it comes to tackling a chore most small businesses waste time on – chasing overdue invoices.

Nadia is the volunteer treasurer of a rowing club in Hammersmith, and uses Sage Copilot to automatically detect when invoices are due – and at the click of a button can send automatic payment reminders to chase unpaid invoices.

Every week, Sage Copilot saves her five hours, she estimates – and Sage customers report that using the function gets them paid a week earlier.

Lisa Ewans, Senior Vice President at the Newcastle-based accounting software firm says: ‘Through our AI, Nadia saves hours of admin a week, which is hugely valuable as a volunteer, because it’s not a day job she’s getting paid for so time is her currency.’

‘Sage Copilot automatically runs in the background, and at the click of a button Sage Copilot chases those customers whose payments are overdue and helps to get the payments in.’

Small businesses can customise the tone of their invoice-chasing emails so it sounds like them, giving them control of how the AI works for them.

Sage Copilot helps small business owners stay on top of their finances and in control of cash flow: as well as chasing invoices. It can help business owners spot duplicate payments and other errors in their accounts, a common problem, and spot trends that might affect the business.

From tradespeople to coffee shop owners

Sage Copilot has a huge amount to offer for businesses in many sectors: tradespeople, for example, often spend considerable time chasing late payments.

Tradespeople can use the time saved to focus on finding new customers or delivering for existing ones.

With Sage Copilot, small business owners can stay on top of their finances with confidence. From chasing invoices to catching duplicate payments and errors, it even spots trends that might affect their business—helping them stay one step ahead.

And when it identifies changes, like rising prices, Sage Copilot provides clear, timely alerts so you can act quickly and make confident decisions.

For other business owners such as cafe owners, the ability to easily spot duplicate payments within VAT returns, or payments that have been miscategorised allows business owners to file returns with peace of mind.

Lisa Ewans says: ‘The focus is around identifying errors in the data, so Sage Copilot spots duplicate transactions for example and alerts business owners automatically. It gives business owners the confidence to sign off returns quickly, and saves them time.’

Sage is a British business, born and bred, with its global headquarters in Newcastle—our home for over 40 years. In that time, it has worked closely with businesses across every sector to understand their challenges and develop financial AI that delivers real value to businesses, solving real world challenges. 

Helping with the details

For accountants and bookkeepers, Sage Copilot offers other benefits, helping them deal with both tax returns and ‘Know Your Customer’ (‘KYC’) checks, which are a series of procedures businesses must follow to verify their customers’ identities and check their risk profiles.

Sage’s experts identified the real ‘pain points’ faced by accountants and bookkeepers, who often spend too much time chasing clients for documents such as identity documents and paper copies of receipts.

Instead of chasing clients via phone and email, Sage delivers automatic assistance when it comes to the documents accountants and bookkeepers need for tax returns and KYC processes.

Ewans says: ‘One of the biggest pain points we hear from accountants is that they spend a load of time chasing up documents,transactions that haven’t been submitted properly, and chasing up paper copies of receipts.’

Sage created a workflow specifically designed so that accountants and bookkeepers could spend less time chasing customers for information.

Instead, Sage Copilot automatically reaches out to customers to request information, and customers upload receipts (for example) as images.

This means that accountants have to spend less time chasing customers via email.

Trusting AI

For Sage, it was important that businesses should be able to trust the AI software to deliver securely, privately and effectively. Sage’s AI Trust Label is a direct response to this – it is designed to provide customers with clear, accessible information about the way AI functions across Sage products.

Sage Copilot was made in the UK, and was built from the ground up with the needs of this country’s businesses in mind.

Ewans says: ‘We know that trust is really important to our customers. You don’t have to be a huge Silicon Valley company to deliver for customers.

‘We have 40 years of experience working with small businesses, and 400 UK-based engineers and data scientists building Sage Copilot to deliver an AI copilot that focuses on the real needs of small businesses today.’