How Tarsadia built its first in-house AI capability with Uplers in 7 weeks

Tarsadia knew the business problem it wanted to solve. The harder part was defining a technical role it had never hired before.

Industry: Private Equity / Venture CapitalHiring challenge: First ever in house engineering hire

But the real challenge wasn’t simply finding an AI engineer.

It was figuring out what the right engineer looked like.

The business shift: Why Tarsadia wanted technical capability in house

Tarsadia Investments is a multibillion dollar private investment firm investing across healthcare, fintech, life sciences, real estate and other sectors.

Its portfolio companies had hired hundreds of engineers over the years.

Tarsadia itself had never hired one.

Historically, technical projects could be handled through external development partners. But the firm was seeing a recurring problem: costs were increasing without a corresponding improvement in quality.

At the same time, AI was changing what a small internal technical team could accomplish.

Instead of relying on a new external development shop every time a technical project emerged, Tarsadia saw an opportunity to bring that capability in house.

A strong AI native engineer could build internal tools, automate investment workflows and support technical initiatives across the firm’s portfolio.

There was just one challenge:

How do you hire for a role your organization has never had before?

The challenge: Defining “great” without an existing benchmark

Tarsadia understood the business outcome it wanted.

Defining the person who could deliver it was harder.

As the hiring lead explained:

“To hire for a position that you don’t have is very hard... my preferences probably evolved as I started having conversations and meeting people.”

3 things made the search particularly difficult:

1. No internal technical benchmark: Without an existing engineering organization, the team couldn’t simply compare candidates against previous successful hires.
2. Evolving requirements: Interviews weren’t only evaluating candidates. They were helping Tarsadia understand which capabilities mattered most.
3. Narrow talent intersection: The role required applied AI and automation skills, financial or investment workflow understanding, and the independence to operate inside a non-traditional engineering environment.

A perfect JD wasn’t going to solve this search.

The hiring process itself had to help define the role.

What Uplers did differently

1. Used the first batch to calibrate the search

Rather than flooding Tarsadia with profiles, Uplers started with a small, tightly curated batch delivered within days of intake.

There were zero early screen rejections in Batch 1.

More importantly, Tarsadia now had real candidates to react to.

Which profiles felt closest? What was missing? Which requirements mattered more than expected?

For a first of its kind role, those conversations provided a more useful signal than continuing to refine an abstract specification.

2. Turned interview feedback into sourcing intelligence

As candidates moved through interviews, feedback was fed directly into subsequent searches.

The process became:

Source → Interview → Learn → Refine → Source again

Instead of treating a rejection as simply a need for another candidate, Uplers used it to understand what the next candidate should look like.

The client noticed the improvement:

“I think the last few batches have improved... The batches got better with each time.”

3. Shifted from volume to staggered precision

As the search became clearer, Uplers moved toward smaller, staggered batches.

That gave Tarsadia time to interview and provide meaningful feedback before the next set of candidates arrived.

Each batch could therefore incorporate what had been learned from the previous one.

The objective wasn’t more profiles. It was a more accurate shortlist.

4. Used AI to search beyond obvious keywords

Traditional sourcing could find candidates with AI experience or financial services backgrounds.

Finding people with both was considerably harder.

Uplers layered AI assisted profile mapping onto traditional sourcing to identify adjacent experiences, capabilities and career patterns that literal keyword searches could miss.

AI expanded the search.

Client feedback and recruiter judgment determined which signals actually mattered.

The results: Tarsadia’s first engineering hire joined in ~7 weeks

What began as an undefined first technical role turned into Tarsadia’s first completed internal engineering hire.

The strongest result wasn’t only that the role closed.

It was that the search became more accurate as Tarsadia learned what it actually needed.

Following the engagement, Tarsadia introduced Uplers to another company in its network, extending the relationship beyond the original search.

The bigger lesson: You don’t need a perfect JD to make a great hire

Tarsadia started with a business problem and a rough idea of the technical capability it needed.

Seven weeks later, its first internal engineering hire had joined.

Not because the original specification was perfect.

Because the search was designed to get smarter.

The process was straightforward:

Understand the business problem → Build a candidate hypothesis → Test it with real profiles → Learn from interviews → Refine the search

For emerging AI roles, that ability to calibrate may matter just as much as the ability to source.

You don’t always need to know exactly who you’re looking for before you start.

You need a hiring process capable of helping you figure it out.

Building a technical capability your company has never hired for before?

Bring us the problem. We’ll help you find the people who can solve it.