Have an account? Log in
How a Silicon Valley AI startup built its engineering team outside the US
How a Menlo Park GenAI company, later acquired by Webflow, hired seven people across nine disciplines without ever running a search in the United States.
159Screening interviews we ran
7Hires across engineering and go-to-market
1 weekFrom first contact to closed, on the fastest hireTL;DR
- Who. Vidoso, a GenAI company in Menlo Park, California. A multi-modal agentic AI platform that turns campaign briefs and long-form video into on-brand marketing assets, used by Samsara, Seismic, Iterable, SmartRecruiters and BILL. Founded by two repeat Silicon Valley founders, backed by Tau Ventures, later acquired by Webflow.
- What they needed. A founding engineering team: senior backend engineers in Python, senior frontend engineers in React. Fully remote, permanent, with real working-hour overlap with California. The scope later widened to machine learning, DevOps, QA, product, UI/UX, sales and customer success.
- Why it was hard. Eight or more years of experience, a founder-led technical interview run by the CTO, and a band of $80,000 to $100,000. That profile is not reachable in the Bay Area at that number, so the search had to happen somewhere else without the bar moving.
- What we did. A written intake per role, sourcing across Latin America, Portugal, Poland and Ukraine, and our own technical interview on every candidate before a name reached the founders. Shortlists in their board within one to two days of a request.
- What happened. Seven hires. Backend, frontend and customer success, plus screening across six more disciplines. One engineer went from first contact to closed in a week.
About Vidoso
Vidoso builds a multi-modal agentic AI platform for B2B marketing teams. Its agents turn a campaign brief into emails, social posts and long-form content, and turn a webinar into clips, chapters and soundbites that can be reused across channels. Customers include Samsara, Seismic, Iterable, SmartRecruiters and BILL. The company is based in Menlo Park, was backed by Tau Ventures, and has since been acquired by Webflow.
It was founded by two repeat operators. Sharad Verma, CEO, previously founded Piqora and BoostUp.ai. Ayush Chordia, CTO, previously founded WrappUp and worked at Cisco and Cruise. Both had exits behind them, and the engineering bar was set by the CTO personally.
They came to us through BoostUp, the founders' previous company, which we were already recruiting for. When Vidoso was starting and the engineering team did not exist yet, they asked us to build it.
The brief
Build a founding engineering team for an early-stage GenAI product, from zero. Senior backend in Python, senior frontend in React, permanent contracts, fully remote, with overlap with California treated as a requirement rather than a preference. Over time the same engagement covered machine learning, DevOps, QA, product, design, sales and customer success. We ran every search from intake to signed offer.
The 4 problems we had to solve
Problem #1: "We need Bay Area engineers on a budget that is not a Bay Area budget."
The backend brief asked for eight or more years of professional experience, Python and ideally Go, large-scale distributed systems, and enough interest in machine learning to be useful around a GenAI product. The band was $80,000 to $100,000. In Menlo Park that profile costs several times that. Hiring in the US was never going to work at that number, and lowering the bar was not on the table.
What we did
We moved the search rather than the requirement. Four markets in parallel: Brazil and wider Latin America, Portugal, Poland and Ukraine. In all four, that band buys a genuinely senior engineer, and a team lead is a realistic target rather than a stretch. One of the backend engineers we placed was working on payments infrastructure at PayPal when we found him. Vidoso never opened a US search for these roles.
Problem #2: "The interview that decides everything is our CTO's."
Vidoso's process ended with a technical interview run by Ayush. It was the stage that actually decided the hire, which meant a partner who only read CVs would have spent the founders' calendars on people who were never going to clear it. With a team that small, that cost is real.
What we did
We put our own technical interview in front of it. The question bank was built with one of our senior engineers, organised by discipline, backend, Python, frontend, React, and banded by years of experience, so an answer is graded against what that level should know rather than against one fixed standard. When the founders wanted something specific probed, we folded their questions into ours. Every question and every answer was written out in the notes we sent over, so the founders could see the reasoning, not just a verdict.
Problem #3: "When we open a role, we need people now."
Requests arrived when a gap appeared, not on a hiring calendar. A shortlist that takes three weeks is a shortlist that arrives after the founders have already worked around the problem.
What we did
We kept candidates screened ahead of demand. Because we were interviewing continuously across their stack, we could open a board with two or three screened candidates within one to two days of a request. That gave the founders something to move on immediately and gave us room to keep sourcing underneath for a better fit. Each card carried the interview notes, salary expectation, location, timezone and availability, so a decision could be made from the card.
Maxwell Dalboni, a backend engineer in Brazil with sixteen years of experience, went from first contact to closed in a week. He had already been through our screening, so there was nothing left to verify.
Problem #4: "We are a ten-person company hiring across nine disciplines."
It started as backend and frontend. It became machine learning, DevOps, QA, product management, UI/UX, sales and customer success. Nine role types, each with a different market, different seniority signals and different questions. A ten-person company does not have the bandwidth to manage a different agency per discipline.
What we did
We ran all of them, one team, one process. Each new discipline got its own intake document and its own screening question set before we contacted anyone, so a QA shortlist was assessed as carefully as a backend one. Vidoso never had to onboard a second supplier or explain the company twice.
How we actually ran it
- One intake per role. Why the hire exists, what the person does daily, who they report to, required against preferred skills, the band, the timezone requirement, and the interview loop with a named owner per step. The backend role was scoped at $80,000 to $100,000, senior, remote from Latin America.
- Sourcing across four markets. Latin America, Portugal, Poland and Ukraine, through our own database and direct outreach. Overlap with California was applied at first contact, not discovered at offer stage.
- Our own technical interview on everyone. Background, motivation, English, notice period, salary expectation, then the technical block from our banded question set for that discipline and level. No exceptions, in any of the nine disciplines.
- A board per role, updated live. One card per candidate with the full interview notes, the technical answers written out, salary, location, availability and contact details. Two to three candidates per role, delivered in one to two days.
- Founder interviews, then a paid trial month. CV review and screening call with Sharad, technical interview with Ayush, then a paid fractional engagement of around 25 hours a week for a month before converting to full-time. Both sides saw real work before anyone committed.
