Quitting a career in venture capital to build a $150k ARR portfolio of health apps

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get The Free Guide

Presh Dineshkumar spent seven years at LAUNCH, Jason Calacanis's venture fund, helping invest in more than 50 startups and watching hundreds of founders build companies from scratch. At some point, watching from the sidelines stopped being enough. He wanted to build something of his own.

He had also been personally obsessed with health for years before that decision. While investing full time at LAUNCH, he spent his own time experimenting on himself, tracking sleep, heart rate variability, cold exposure, sunlight, and nearly every biomarker he could get his hands on. He noticed a gap: plenty of tools existed for collecting health data, but very few products actually helped people change their behavior day to day.

That gap became The Wellness Company, the entity behind Tempo, along with a small portfolio of single habit apps: GoPolar for cold exposure, SunSeek for sunlight, and Posture AI. Each app started narrow, focused on one daily habit. As people began paying for them, a bigger pattern emerged. Users did not actually want another standalone health app. They wanted one place that understood all of their health data together and helped them make better decisions with it.

Today the portfolio generates more than $150k in annual recurring revenue from consumer subscriptions, entirely bootstrapped. The individual apps are now converging under Tempo, with the long term vision of becoming a single AI powered health companion that combines wearable data, biomarkers, and everyday habits into one place.

Here is how a venture investor built a bootstrapped health app portfolio, and why he now believes distribution matters more than most founders want to admit.

The first version of the product was intentionally simple. Presh and his cofounder built on nights and weekends while he was still working in venture capital. Rather than attempting a full health platform from day one, they focused on shipping small products, talking to users, and learning as fast as they could. The company bootstrapped by launching individual health apps and reinvesting every dollar of subscription revenue straight back into the business.

Becoming part of the local builder community had a disproportionate impact on how the company actually got built. Presh attended founder meetups and hack nights not to network for fundraising, but to learn what other builders were actually using. That is how he first discovered Supabase, years before it became a core part of the backend. Those conversations often saved weeks or months of research, because the lessons came directly from people who had already solved the same problems.

Once the company had paying customers and steady subscription revenue, that traction gave enough confidence to raise a small angel round, used to hire a small team, speed up development, and expand the app portfolio. Presh is direct that funding, in his view, exists to amplify something that is already working, not to go searching for product market fit in the first place.

That same mindset has only become more important with AI moving as fast as it has. The landscape shifts almost weekly, and reading blog posts is not the fastest way to stay current. Talking directly to founders who are actively building is. New models, coding tools, infrastructure, and workflows consistently surface inside the builder community long before they become mainstream knowledge.

Every product starts in Figma, using AI and the Figma MCP to rapidly explore ideas and iterate on the user experience before a single line of code gets written. That lets the team test flows and understand how something should work without spending days building the wrong version first.

From there, the apps get built in Xcode using AI coding agents like Codex and Claude Code. Presh does not have a traditional software engineering background and describes himself as a non-technical founder, but AI has genuinely changed what that label means in practice. He can now build and ship production quality iOS features alongside the engineering team by working with coding agents rather than writing every line himself.

One habit he particularly likes: running Codex on a dedicated Mac mini connected to the GitHub repository around the clock. That setup lets him review code, fix bugs, merge pull requests, and even ship production updates from his phone while away from his desk entirely.

The backend leans heavily on Supabase. That combination, Figma for design, AI coding agents for implementation, Xcode for the platform, Supabase for the backend, lets a very small team build and maintain several consumer health apps with far more leverage than would have been realistic just a few years ago.

One of the biggest technical challenges was migrating away from Apple's CloudKit. It worked fine in the earliest days, but as the products and backend grew more sophisticated, the team needed the flexibility and portability that Supabase offered instead.

The migration itself was not simple, since it required supporting both systems simultaneously while moving users over without disrupting anyone's experience. The team built custom migration tools that synced data between CloudKit and Supabase, allowing an incremental transition rather than a single risky cutover.

AI coding agents made a real difference throughout that process. Instead of spending weeks writing internal tooling by hand, the team quickly built migration scripts, validation tools, and utilities that automated much of the work. What could have been a genuinely painful infrastructure project became manageable, and it reinforced a broader belief: AI is especially powerful for exactly the kind of engineering work users never see, but that every company eventually has to deal with.

The business model is straightforward. Users pay monthly or annually for app access, and nearly all revenue comes directly through the App Store.

Pricing has changed several times as the team learned more about customer intent. Superwall is used to test different paywalls, pricing structures, and offers, varied by onboarding flow and user segment. One consistent finding: hard paywalls can work extremely well when the product solves a clear, immediate pain point. Founders often hesitate to ask users to pay too early, but if the value is obvious, people are willing to pay for it right away.

Revenue expansion runs through several levers at once: improving conversion, increasing annual plan adoption, reducing churn, and launching additional products aimed at the same health conscious customer. Each standalone app also creates a natural path into Tempo, which can eventually become the broader subscription that pulls a user's health data, habits, and personalized recommendations into one place.

In the early days, distribution was treated exactly like product development: a series of small, deliberate experiments. Before building something, the team often validated demand by posting ideas on Reddit or X, building simple landing pages, or shipping free tools that solved a small, adjacent problem. Before expanding Tempo, for instance, the team launched a free AI Body Fat Calculator on the web, which generated awareness and introduced people to the broader product.

Today, the biggest growth channel is short form content on TikTok and Instagram. Rather than chasing creators with the largest audiences, the team looks for people whose content consistently earns strong engagement. The usual process starts with finding one creator who matches the ideal customer profile, then using platform recommendations to surface dozens of similar creators. From there, they reach out directly, send the product, and let each creator make content in their own style.

That single result changed how the whole team thought about creator marketing going forward. Every meaningful feature launch is now also treated as a chance to relaunch the company. Each significant release includes a short founder led video, usually under 90 seconds, showing the product, explaining the problem it solves, and sharing it across X, TikTok, LinkedIn, and other communities. Consistently putting the product back out into the world has created opportunities that could never have been planned for, from customer feedback to partnerships and introductions that showed up simply because someone saw a launch post.

For consumer apps specifically, content is currently one of the most effective distribution channels available, because it lets you demonstrate both the problem and the product in a way that feels native to how people already discover new tools. The biggest lesson from all of it is that distribution compounds. Building an audience, creating content consistently, working with creators, and launching every new feature publicly has been far more valuable than waiting for the product to feel finished before saying anything about it.

Customer acquisition has been the single biggest challenge throughout. Building products has become dramatically easier with AI, but getting people to actually discover, trust, and pay for a product remains genuinely difficult. A significant amount of early time went into improving the product when more of it should have gone into improving distribution instead.

Starting over today, Presh says he would think about distribution before writing a single line of code: where does the customer already spend time, what kind of content naturally reaches them, who already has their attention, and how can demand be validated before anything gets built.

He would also not wait until launch to start talking about the product. Building in public from day one, creating content throughout the process, growing an audience alongside the product itself, and treating every feature as another opportunity to launch again would all start immediately rather than after the fact. Distribution is not something bolted on once a product is finished. It is part of building the product in the first place.

The biggest mindset shift for Presh personally has been accepting that the best product does not always win. The product that reaches the most people usually does.

A handful of influences have had an outsized effect on how Presh builds. On the product side, companies like Apple, Linear, and Figma inspire him constantly, all of them sharing an obsession with craftsmanship, simplicity, and sweating small details. He estimates he spends nearly as much time studying great products as he does building his own.

He also spends a lot of time on X, where he discovers new tools, learns how other founders are solving problems, and stays current on how AI is reshaping software development. Thinkers like Naval have shaped how he thinks about building companies generally, but he learns the most, day to day, from following builders who openly share exactly what they are shipping each week.

Outside of work, a few simple habits shape his process. Most mornings start with a cold shower, which helps him wake up and focus. He prioritizes walking 7,000 to 10,000 steps every day, and those walks function as real thinking time. Many of the team's best product ideas, design decisions, and solutions to difficult problems have emerged on a walk rather than in front of a computer.

Embracing AI early was, in his own assessment, one of the best decisions the company made. It dramatically increased the small team's leverage and let them move considerably faster than would otherwise have been possible. At this point, it is difficult for him to imagine building software any other way.

The long term goal is to build the world's best AI health companion, built on the belief that healthcare should be proactive rather than reactive, and that everyone deserves access to personalized guidance based on their own wearable data, biomarkers, and daily habits.

In the near term, the focus is growing Tempo into a product that brings together everything learned from building GoPolar, SunSeek, and Posture AI: one place that helps people understand their health, tells them what to do next, and makes staying consistent easier.

Beyond software, the team is also excited about building hardware that integrates seamlessly into the Tempo ecosystem, believing that AI, software, and thoughtfully designed health devices together will unlock entirely new ways for people to understand and improve their own health.

The mission, at its core, is simple: help millions of people live longer, healthier lives by making healthy decisions easier every single day.

You can follow along on X, where Presh is most active and would love to connect with anyone building something interesting, or on Instagram. You can also join a step challenge directly through iMessage.