Growing a talent marketplace to $2.5M+/yr after two years of losses

Better prompts. Better outcomes.

If you're still typing the same basic requests into ChatGPT, you're leaving productivity on the table. One great prompt can replace hours of manual work.

Access 100 ready-to-use ChatGPT prompts designed to help you automate tasks, generate ideas, solve problems faster, and boost your daily productivity.

David Stepania was born in the Republic of Georgia as the Soviet Union was collapsing, attended 13 different schools across multiple countries growing up, and eventually landed in Seattle, where he studied at the University of Washington. That upbringing taught him one lesson above all others: resourcefulness. There were periods when textbooks were a luxury, so "make it work" became his default mode of operating, which turned out to be excellent training for founding a company.

Today he runs ThirstySprout, a talent marketplace that places remote AI and engineering talent, mostly from Latin America, Eastern Europe, and Asia, with funded US startups and enterprises. Companies including The Real Real, Hopper, Mailchimp under Intuit, Rover.com, and a wave of VC-backed startups have used it to build their technical teams. He also built ChoppingBlock.ai, an AI salary and jobs intelligence platform with a newsletter and podcast that has grown past 40,000 subscribers and doubles as the company's audience engine.

ThirstySprout crossed $2.5M in annual revenue a while back, and the growth since then put it at number 247 on the Inc. 5000 in 2024. This year, the goal is to double that and push past $5M, with a team that is dramatically leaner than it was the first time around.

Here is the full arc, from a Hawaiian beach and two years of credit card debt to a lean, AI-native operation with real margins.

ThirstySprout came directly out of failure. David was coming off a sabbatical in Hawaii after a previous venture, nursing a few product startups that had not worked out. The realization that hit him on that beach was simple: to build an extraordinary startup, you need an extraordinary technical team, and finding one you can trust is brutally hard. He did not have a product idea, so he decided to solve the problem he already understood firsthand.

The company started as a two-person product development agency, and the early years were a catalog of classic mistakes. They took on underbudgeted projects and over-delivered just to earn testimonials. After enough of that, it became clear the model did not scale, and a friend suggested they focus on staffing remote technical talent instead.

The pivot validated itself almost immediately. A single cold email landed Rover.com as a client during its hypergrowth phase in 2018. A six-figure engagement and a reference from a hot startup changed everything, and suddenly the team could win clients of that caliber repeatedly.

The financial picture behind that early traction was rough. David had close to zero in cash reserves and funded the gap on personal credit cards, running the business at a loss for the first year or two, peaking somewhere between $50k and $100k in debt, while paying himself no salary. He does not recommend that path, but he is direct about the upside: stubbornness turned out to be a real business asset.

For years, the product was a service run entirely on top of other people's tools. No custom platform, no code. The team deliberately refused to build technology until they had 50 to 100 actively engaged freelancers and proven, repeatable processes, on the reasoning that automating a process you have not proven by hand just scales your mistakes.

The initial version ran on WordPress, then low-code tools like Webflow. The real build was on the supply side: the team went country by country looking for deep talent pools, and found their strongest footing in Georgia, the country David was born in. The cultural fit and engineering work ethic were exactly right, and today ThirstySprout is one of the top talent players there, competing directly with Toptal and Turing.

The initial build cost no cash. It cost two loss-making years on credit cards, and the time to learn an industry David had never operated in. The people who helped most were early clients who took a chance on the young company, and the founder communities he leaned on for advice along the way.

Today, the spine of the company is Claude, and David means that literally rather than as a buzzword. Candidate screening and ranking, financial analysis, contract drafting, content, and strategy mostly run through dedicated Claude workspaces, one per business domain, each loaded with custom instructions. For agentic builds, he uses Claude Code.

Around that core:

The biggest stack change of the past year was subtraction, not addition. Earlier this year, David audited the company's subscriptions and found more than 50 tools, including five overlapping AI subscriptions. The team cut aggressively. His rule going forward: one tool used at 90% depth beats five used at 10%.

ThirstySprout runs on two fee models, and the mix between them reflects the business's current shape:

Historically, staffing engineers made up the bulk of the business rather than direct placements, though today the split is close to even. The company has charged from day one; as a bootstrapped operation, it never had a free-work phase to graduate out of.

The model includes natural expansion in two directions. Land-and-expand is the first: a client who hires one engineer and has a good experience often hires more, and most of the company's revenue now comes from repeat placements rather than new logos. Contract and hourly engagements are the second, giving clients flexible capacity between placement fees and smoothing revenue in the process.

A bigger expansion play is being built around the placement business itself. ChoppingBlock functions as an audience and data asset, with salary data, jobs intelligence, and newsletter sponsorship potential feeding the supply side, while the marketplace product reduces cost per placement. The recruiting fees fund the flywheel, and the flywheel makes each future fee cheaper to earn.

For years, ThirstySprout grew primarily through high-volume cold email. Early on, the team sent a few hundred emails daily with open rates in the high 50s. At its peak, volume climbed to thousands of emails a day, but open rates dropped to the low 20s and reply rates fell to 2 to 3%, a clear signal of the channel's declining effectiveness. It worked. It built the company. But David considers it dying: AI-generated outreach will soon bury every inbox, and no one will be able to out-automate spam.

Over the last two years, the team deliberately rebuilt its growth engine around inbound, through three channels:

His conclusion is the strategic core of the whole story: rented channels like cold email got the company to millions in revenue, but an owned audience is the only durable moat in a world where AI makes outreach essentially free.

The early challenge at ThirstySprout was survival: two years operating at a loss, funded by personal credit cards, no salary, in an industry David had never worked in before, with underpriced projects, no niche, and no systems.

The mid-stage challenge was self-inflicted. Growth started to feel like headcount, so the team scaled to more than 50 contractors and drifted away from the lean operation that had made the business work in the first place. Unwinding that, and getting honest that a smaller, AI-augmented team could produce more than a bigger one, was both operationally and emotionally difficult.

Three deliberate changes followed:

The company now runs dramatically leaner and produces more per person than it did at peak headcount. The current challenge is the industry itself: AI agents are commoditizing exactly the part of recruiting that most firms live on, namely volume sourcing, resume matching, and mass outreach. David's response has been to deliberately abandon that layer and move up to the judgment layer instead: deep vetting, fit, and retention.

He is equally direct about timing: he would start building an owned audience from day one instead of renting attention through cold email, and he would set a 6 to 12 month window to test hypotheses instead of letting stubbornness fund losses on credit cards for two years.

Books shaped the foundation of David's thinking. He has listened to more than 200 on Audible, working through most of the standard business and self-help canon, including Zero to One, Deep Work, Essentialism, and Principles. His honest takeaway is that many ideas repeat across books, and you often read an entire book to find the one moment that unblocks you. But that single moment can be a genuine turning point, and sometimes an idea he picked up five or ten years ago clicks precisely when he is stuck today. He cannot name one book that changed everything; the value is cumulative, a library to draw on at the right moment.

The biggest resource shift in his career, though, happened recently. He stopped reading about leverage and started using it. AI, Claude specifically, has been the single greatest advantage of his founder life. He performed CFO-level financial analysis that surfaced real margin issues in the company's books, rebuilt the entire legal stack including the master service agreement and contractor agreements with AI drafting and a lawyer reviewing, and built an AI coach inside Claude Code that pressure-tests his decisions before he commits to them. As a solo CEO, he calls it the closest thing to a cofounder he has found.

He pairs that enthusiasm with a hard rule: never automate a process you have not proven manually. Every process at ThirstySprout was scaled by first doing it manually, painfully, until the team truly understood it. Every automation disaster the company has had came from skipping that step.

The headline goal is to grow ThirstySprout past $10M in annual revenue, and to do it the way it has always been done: bootstrapped, with a lean AI-native team, rather than a bloated one. Beyond that number, there are two tracks.

For ThirstySprout, the plan is to productize the marketplace, moving from a services company using software into a software-enabled marketplace with service margins, without taking VC money. The AI-native talent niche is expanding quickly, and the intention is to own it.

For ChoppingBlock, the goal is to become the Levels.fyi of the AI era, the default place people check to see what AI-era roles pay and which skills are rising or dying. The flywheel runs on user-submitted compensation data unlocking aggregate insights, wrapped in a newsletter and podcast people genuinely enjoy.

There is a personal goal too. David is currently working from the beaches of the Riviera Maya near Cancun, which he is not complaining about, but the real target is to someday run all of this from Hawaii. The idea for ThirstySprout came to him there, so building the company into something that lets him work from that same beach would close the loop, which means a great deal to him.

The overarching goal, in his words: prove that he can build a modern, AI-native company of real consequence with a tiny team and zero outside capital.

You can follow along on X, LinkedIn, and his personal site, or check out The AI Chopping Block podcast.