How Startup Leaders Are Recruiting and Upskilling for the AI Era
- Stephanie Roulic
- Jun 8
- 6 min read
At Startup Boston Week, the panel “Closing the Gap: Recruiting and Upskilling for AI Success” tackled one of the startup ecosystem’s biggest anxieties head-on: how workers, founders, and companies can adapt to a world increasingly shaped by AI.
Moderated by Rizel Scarlett, Tech Lead of Open Source Developer Relations, Block, Inc., the conversation brought together leaders across engineering, recruiting, and AI startups, including Luisa Herrmann, founder of AINova; Catherine Weeks, Director of Engineering at Red Hat; and Tommy Barth, Head of Recruiting Operations and Analytics at Apollo.
While much of the public conversation around AI centers on fear - job displacement, automation, and hiring disruption - the panel focused on something more nuanced: what it actually means to become “AI-ready.”
“There seems to be this pressure to learn AI or get left behind,” Scarlett told the audience early in the discussion. “But people might be like, what exactly is AI?”
AI Is Bigger Than ChatGPT
One of the panel’s first goals was clarifying what people even mean when they say “AI.”
Barth pushed back against treating AI as a catch-all buzzword.
“A lot of people talk about AI as an umbrella term,” he said. “It’s like saying, ‘I did it on the internet.’ Like, where?”
The panelists distinguished between predictive AI, systems that have existed for years powering things like fraud detection, spam filters, and navigation apps, and newer generative AI systems powered by large language models.
“Technically a calculator is AI,” Herrmann joked, before explaining that AI has quietly existed in everyday tools long before the recent explosion of ChatGPT and Claude.
Weeks described the current moment as a shift from deterministic systems into more creative, non-deterministic tools.
“LLMs brought us into a non-deterministic world where we can create with AI,” she explained.
But she also pointed toward what she sees as the next major frontier: AI agents and MCPs, or systems capable of taking action on behalf of users rather than simply generating information.
“How do you connect and allow AI to actually take action on your behalf rather than it just giving you information?” Weeks asked.
The New Baseline for Technical Talent
As the discussion turned toward hiring, the panelists made clear that AI fluency is rapidly becoming an expectation rather than a differentiator.
Barth explained that at Apollo, AI competency is already appearing in interview processes and performance reviews.
“We now have AI interviews that are specifically focused on ensuring that engineers and anyone who’s in a technical role has not just an interest in, but like a level of fluency in AI,” he said.
He added that employees are increasingly expected to explain how they’re using AI tools to improve efficiency in their day-to-day work.
“The writing is on the wall here,” Barth said.
Still, the panel repeatedly emphasized that “AI-ready” does not mean every engineer needs to become a machine learning researcher.
“You don’t have to be focused on it,” Herrmann said. “You just need to know how it works and work with it.”
She cautioned against developers treating avoidance of AI as a badge of honor.
“That’s not the flex you think it is,” Herrmann said.
Instead, the panel framed AI as a tool that can eliminate repetitive or low-value work, allowing employees to focus more energy on higher-level thinking and execution.
“If it’s a repetitive task, if it’s something that I don’t want to dedicate my brain power to, I will offload that task,” Barth said earlier in the discussion.
The Fear of “AI Slop”
Even as the panel advocated for AI adoption, several speakers warned against over-reliance on generative tools.
Barth described the growing flood of low-quality AI-generated content as “slop.”
“You’ve not actually done anything,” he said. “You’ve just created more information for us to all consume.”
The conversation eventually turned philosophical, with Herrmann introducing the audience to the “dead internet theory,” the idea that one day the internet could become dominated entirely by bots generating content for other bots to consume.
“At some point all the content on the internet will be produced by bots and all of the content will be consumed by bots,” Herrmann explained.
The panelists agreed that the real value will increasingly come from people who know how to ask thoughtful questions, validate outputs, and apply judgment rather than blindly accepting AI-generated work.
“You have to be a world-class product manager to know the right question to ask,” Barth said while discussing prompt engineering.
Weeks added that engineering teams still need to maintain quality standards, even if AI accelerates development.
“If you’re doing stuff that wrecks the quality or wrecks trust, then you’re doing it wrong,” she said.
Why Diversity in AI Is a Business Problem
One of the strongest moments of the panel came during an audience question about bias in recruiting and AI systems.
Herrmann argued that the lack of diversity in AI development is not simply a moral issue, it is a product problem.
“If you have people developing software who all look the same, sound the same, and think the same, that software is going to work for them and not for anybody else,” she said.
She pointed to historical examples of biased technology, including cameras that failed to properly detect whether certain users’ eyes were open because teams had not considered how differently people look across ethnicities.
“There is a need to bring more people in for the reasons of creating better products,” Herrmann said.
Barth took a similarly strong stance when discussing AI in recruiting.
“It’s so important that AI is not making decisions for you. Period,” he said.
He explained that while automation can help recruiters process large application volumes, hiring decisions still require human judgment.
“We will probably live during a time when there are companies that are making all of their hiring decisions with AI,” Barth said. “I won’t work there.”
At the same time, he noted that AI tools can help expose human bias rather than reinforce it. Barth described using interview analysis software that flags when interviewers interrupt candidates too often or when certain groups receive less speaking time during interviews.
The AI Productivity Paradox
Audience questions also pushed the panel to address burnout and rising workplace expectations.
One attendee questioned whether AI productivity gains would ultimately just lead companies to demand even more output from employees.
Weeks acknowledged that many companies are currently operating in what she called the “AI rat race,” where organizations are aggressively experimenting in order to stay competitive.
Still, she argued that long-term sustainability remains critical.
“At the end of the day, we’re still approaching it as this is a 40-hour work week and we want it to be a 40-hour work week,” Weeks said.
Herrmann suggested the AI era may also accelerate more flexible work arrangements, including fractional or project-based expertise rather than traditional full-time structures. “Now their brain power is exacerbated,” she said, referring to employees augmented by AI tools. “And now I’m supposed to expect more from them?”
Recruiting Is Becoming More Human Again
Despite the heavy focus on automation, one surprising theme emerged repeatedly throughout the session: the growing importance of in-person relationships.
Barth told attendees that networking and direct human connection have become more valuable, not less, in the AI hiring era.
“I cannot understate the value of in-person and people-to-people connections in this job market,” he said.
He encouraged job seekers to proactively reach out to professionals they admire, schedule short conversations, and treat networking like long-term business development.
“Don’t give up on building relationships and using your network,” Barth said.
Herrmann echoed that point while reflecting on her own recent hiring experience after receiving over a thousand applications for a role.
“I ended up having to talk to people and say, ‘Hey, I just got way too much. Can you help me find someone?’” she said.
The result, she admitted, was that personal connections ultimately mattered more than application portals.
“I think we’re coming back to in-person interactions that are most important for hiring,” Herrmann said.
Scarlett added that “learning in public,” posting projects, sharing experiments, and documenting growth online, can also help candidates stand out in increasingly crowded markets.
“Nobody Has All the Answers Yet”
As the panel wrapped up, Weeks left the audience with a reminder that the AI transition is still unfolding in real time.
“Nobody has all the answers,” she said. “It’s a time of experimentation.”
Rather than waiting for certainty, the panel encouraged attendees to stay curious, experiment often, and focus on learning how to work alongside AI rather than competing against it.
For many in the room, the message was less about mastering a specific tool and more about adapting to a rapidly changing way of working.
“The most important thing to know right now,” Weeks said, “is that nobody has all the answers.”
