It’s no surprise that many of the best AI researchers on the job market have their eyes on giants like OpenAI or Anthropic, or whatever latest model is making headlines. The draw of joining an established and prestigious company is hard to resist.
To stay competitive in today’s market, software companies like mine need to take AI seriously. That means we need to go head-to-head with the most iconic AI companies on talent. This feels daunting, if not impossible, but it’s something we’re actively trying to get better at.
Here are four steps we take to find and secure the best AI talent, despite being a small business.
What Is an AI Researcher?
An AI researcher researches model architectures and inference efficiency/optimization and trains new models or fine-tunes existing models for a particular purpose. There’s a level of innovation to an AI researcher that goes beyond using existing models to build products — they create something fundamentally new or improve something at the model level.
Round 1: Highlight the Kind of Problems the Candidate Will Get to Solve
Great technical talent loves to solve hard problems. With that in mind, our first step in the interview process is to delve into specific challenges they’ll get to work on within the business in their role, as well as the impact that’s possible if they solve the challenge.
We frame these challenges at the top in order to get a candidate thinking and give them the best possible look into the role: they’ll be working on finding new ways to use AI to reach our customers and move the needle for their businesses while maintaining brand voice and consistency.
At Postscript, we have a treasure trove of data from real customer interactions. The challenge? Turning it into AI-powered conversations that translate to return on investment for the brands we work with. By highlighting the candidate’s opportunity to solve technical problems on the job as early as possible, we invite potential hires to imagine themselves tackling complex issues that can make a tangible difference.
Round 1 Part 2: Present a Bespoke Technical Test
We have a two-part recruiter screen with our engineers. Next, we conduct a technical test to determine a fit.
Early in the recruiting process, we’re getting a sense of their skills. When the candidates move further into the interview rounds, we present them with complex problems that we’re solving and actually speak to what’s happening at our company: interesting, deeper and more technical questions and ideas that show them what they would actually be doing (not hypotheticals).
When potential team members understand that they could be instrumental in shaping a product from the ground up — rather than just fine-tuning an existing system — it can be a game changer, especially for the particularly entrepreneurial-minded candidates.
Round 2: Lean Into the Creative Freedom at Small Companies
A lot of the best technical talent want to be able to share their perspective and explore new ideas with their team. This is where being a smaller company can work to your advantage.
With fewer decision-makers and less red tape, things move faster. You can test ideas in real-time, and ship products in weeks or months rather than years. Candidates often appreciate knowing they can make bold moves and take risks in a way that isn’t possible within larger companies.
Similarly, we showcase how much more cross-functional work you can do at a startup in comparison to large companies with rigid structure and hierarchy.
AI researchers can often feel like they’re in a silo, plugging away at the task at hand without visibility into other parts of the organization. We make it clear that this won’t be the case if they join a startup like ours.
Despite being a startup, we’re using the best tools on the market to support our engineers, and we share that information with them up front. We’re solving problems as a company, not just as an engineering team, although engineering is at the core of that. As a result, candidates have the opportunity to see that they’ll be part of a cross-functional team where everyone values each other’s ideas, regardless of where they come from.
Round 3: Determine a Value Fit
Our final round, whether engineering or customer support or sales, is all about core values. We call it our Fearlessness, Excellence, Animal, Customer First and Humility interview (or “FEACH” interview). It’s not just whether you’ll fit in. It’s about how you’ll move the company forward: we feel your value and you feel ours. Our tenured team will conduct this interview to determine whether individuals can make the company better than they found it.
Attracting top AI talent is about showing where we’re different: the unique opportunities and impacts that come with working at a startup and drawing innovative thinkers to our teams. We want to create an environment where employees feel they can truly make a difference and that they’re not just another cog in the machine. We want them to take risks, build from scratch and collaborate. It’s a win-win for both the company and our employees.