Arpit Tripathi

The Engineer Who Chose the Other Door

The offer on the table out of college was hands-on development, the traditional path for a computer science engineer. Then a staffing company came to campus. Tripathi read up on the industry, thought about what it actually did, and made a decision that surprised almost everyone around him. "Instead of going the traditional developing and software engineering side, let me try this," he recalls thinking. "See if it works out."

It worked out. He joined Collabora and found himself managing an account where he performed well quickly, then leaned into the technical work he genuinely enjoyed: core networking and infrastructure hiring, the kind of requisitions that punish recruiters who cannot hold a real conversation about the stack. Three and a half years in, he was among the firm's top performers.

The turn came when he was handed a brand new client, one of the top banks in the world. The team was new and, by his read, short on leadership. He did not wait for a title. "I thought, it's a new team, I think I can share my experience," he says. That instinct, to lead the work before anyone formally assigned it to him, became the pattern for everything that followed. He moved from talent acquisition manager to director, and since 2016 has collected five Talent Engine Leader of the Year awards at his company, including two back to back and one in 2020. He still points to 2018 as his best year in the space.

Ten Years, One Mentor

Ask Tripathi who shaped him and the answer is immediate and singular: Saqib Khan, his boss from 2016 to today and himself a Talent 100 honoree last year. They met once in Detroit while Tripathi was working a different account. Nothing was the same afterward. "Ever since, things have changed and we never looked back," he says.

What Tripathi took from Khan was not a technique but a posture. He watched a leader who started as a recruiter and became a vice president by constantly evolving against market conditions rather than defending a familiar playbook. And he learned what real sponsorship feels like from the receiving end. Building an AI transformation program inside a talent function requires backing from above, and Khan supplied it. "He gives me a free space, honestly, to just do things in my own way," Tripathi says. It is the reason a ten year tenure at one company reads as an ascent rather than a plateau, and it is a standard he now carries into how he manages the leaders coming up beneath him.

Twelve Agents and a Workshop

Ascendion is an end to end software engineering company with its own platform, AVA, which gives Tripathi something most talent leaders never get: a build environment, with the option to route a given task to OpenAI models or to Sonnet depending on what the job requires. He used it. Today his team runs roughly twelve working agents in production, deployed not only to his own group but to recruiters across the globe.

The first one came out of a problem everyone in the industry recognizes and few had addressed. Every resume is now beautifully written. Every job description is now beautifully written. The traditional matching layer, AI comparing an AI resume against an AI job description and returning a tidy ninety percent match, had quietly stopped meaning very much. Tripathi built an agent that returns a guideline instead of a verdict, flagging that a document may be, say, sixty percent AI written so that the recruiter knows what to probe in screening and technical interviews. He is careful to say this is not a judgment. "I'm not saying AI-written resumes are bad," he notes. Everyone is using the tools. The point is to see clearly.

Then came the harder problem of identity fraud and inflated profiles. His team built a technical consistency agent that surfaces claims that do not hold together: a junior developer who appears as a network architect in the very next role, or a candidate credited with fourteen years on SharePoint 2018. Other agents handle end to end outreach with customized templates, auto triggers and interview messaging, reformat candidate summaries into the specific presentation format certain clients demand, and route a strong candidate toward the client and industry where their background will actually land rather than letting a good profile die in the wrong pipeline.

Just as telling is how he scaled it. Rather than handing recruiters finished tools, he ran a hands-on workshop with a simple deal: find your own use case, build the agent, show it to me. If you cannot build it, bring me the use case and we will build it and launch it for everyone. That is how a personal experiment became an organizational capability.

Why the Recruiter Stays

For someone this deep in automation, Tripathi is unambiguous about its limits. Asked constantly whether agents will replace recruiters, his answer does not waver. "No, absolutely not," he says. Technical screening still demands human attention, and even after an AI interview he wants a recruiter validating the outcome. The goal of every tool he has shipped is to buy back the hours so that people can spend them where a human in the loop genuinely changes the result.

He is equally generous about the pace of adoption. Some people are ready to accept the change immediately; others take it slowly, which he considers fair, because everyone learns differently. He knows the cost of that transition firsthand. Moving veteran recruiters off a method they have used for ten years takes repeated conversations, persistence and, occasionally, a standoff. He has lived through it, and he does not pretend it is easy.

Match the Pace

What excites him most right now is the newest cohort. Recruiters arriving today never had to unlearn anything. They studied alongside these tools, they already expect matching and scoring to be handled for them, and they skip the change management everyone else is grinding through. Tripathi has watched them move genuinely fast.

His advice to them is built around a company philosophy he clearly believes in: the 10X program, the idea that regardless of tenure a single person can do the work of ten as long as they use the right AI tools. So the counsel is two lines long. Keep doing what you are already doing, because arriving fluent in AI is an advantage, not a shortcut. And match the pace of the current flow as fast as possible.

It is the same instruction he has been following since he walked away from a developer role to try something no one expected. The industry keeps changing shape. Tripathi keeps building the thing that fits it.