Jason Kirby: Welcome back to the show everyone. Today we have Pratik Josie with— Jo… I said Josie. Right. I’m just going to cut that. All right, everyone, welcome back to the show. Today we have Prateek Joshi with us, founder and CEO of Plutoshift. Welcome to the show.
Prateek Joshi: Jason, thank you for inviting me. It’s great to be here.
Jason Kirby: I’m excited to have you. You have a fascinating story as an exited founder, a machine learning author who’s written 13 books—and not little books, like textbooks. Before the days of LLMs where you could have ChatGPT do it for you. You built a great company and now you’re a VC. How did you get here?
Prateek Joshi: During my 13-books phase, I wish I had LLMs—could’ve really used that. I wrote them before all of this. Quick story: I was born and raised in southern India, in a small town. I came to the US for my master’s in LA, then moved to the Bay Area for my first job at Nvidia on the AI team. Back then it was very early—could we make it work on mobile devices? Do we have enough compute? Will applications be useful or just demo‑ware? I’ve been building and shipping AI products for a while.
Prateek Joshi: I also wrote a blog—if I discovered something useful, I’d write about it. One day a publisher reached out: “We like your blog. Want to write a book?” I thought, how hard can it be? I’ve written blog posts. The first book was excruciating—300 pages with rounds of review. I got through it, kept writing, and after 13 books I told the publisher, I’ve got nothing else to say—I’m out.
Jason Kirby: You got to 13 books—substantial, hundreds of pages each on Python, AI, and machine learning. For reference, you were at Nvidia around 2010–2012, then wrote the books for a couple years. You were at the forefront of AI and ML. How did that lead to starting a company?
Prateek Joshi: I see AI as the goal and machine learning as a vehicle to get there—data is the fuel. Writing the books showed me how much you need to know to make ML practical—moving from theory to the real world. I wanted to bring AI into the physical world—beyond images and search—to infrastructure. That insight led to launching Plutoshift. We started with water infrastructure—close to my heart because I grew up where water was scarce—then expanded to other physical infrastructure. We served Fortune 500 customers for about seven years.
Jason Kirby: You raised over $16M for Plutoshift and you were a solo founder. What was it like raising capital? What came first—company or capital?
Prateek Joshi: Once I decided this could be a company, I did customer research—would anyone care? I found a small group extremely enthusiastic—that was the first signal. The biggest barrier was visa status. I didn’t have the resources for a personal lawyer. Most investors said, “Figure it out then come back.” I met Unshackled Ventures—they specialize in this and wrote the first check. That pre‑seed was around $300k with a couple of angels—to validate the hypothesis and build a real product and pilots.
Jason Kirby: How much did you raise in the seed and how did you get in front of those investors?
Prateek Joshi: The visa was sorted by pre‑seed close—you can’t really raise otherwise. For seed we raised about $3.5M—it started as ~$2M and expanded with interest. My network was zero—no friends‑and‑family. I did cold outbound: LinkedIn, websites, events. Some good people helped with introductions. That’s how I raised.
Jason Kirby: Let’s talk Series A. What metrics did you show and how much did you raise?
Prateek Joshi: We’re enterprise software—selling to big companies. Metrics: number of paying customers (logos), overall revenue and median ACV, and expansion (deploying to a second facility, more users, features). At Series A we had north of 10 large logos with signed contracts, late‑five‑figure to early‑six‑figure ACVs, and expansion in roughly a third of accounts. Investors did customer reference checks—I gave them phone numbers.
Jason Kirby: Fundraising difficulty by stage?
Prateek Joshi: Pre‑seed was hardest because I knew nothing about visas, fundraising, or company building. Series A was also hard—enterprises move in quarters while startups live in days. Seed was the easiest by comparison.
Jason Kirby: You ultimately sold the business. Why was that the right move?
Prateek Joshi: It was an asset acquisition. The company continues under new management—I’m friends with the new CEO. After seven years, leadership and investors wanted to pursue different next chapters. There was a path for the product and name to live on, potentially in a different direction, so we took it.
Jason Kirby: Now you’re on the investor side at Moxie Ventures. Why VC and what does your day‑to‑day look like?
Prateek Joshi: In the final third of my founder journey, early founders kept reaching out for help. I connected great founders to the right investors and wished I could write a check. I enjoyed it, so after wrapping up Plutoshift I moved to VC full‑time. Now I focus on sourcing great founders, researching subsectors (vertical AI, robotics, climate, healthcare, AI+bio), customer sentiment work, running a repeatable diligence engine, winning competitive deals, and helping portfolio founders go from Seed to A.
Jason Kirby: What types of deals does Moxie invest in—stage, sectors, check size?
Prateek Joshi: We’re a seed‑stage fund, typically when a company is raising $2–4M; we’ll often write the largest check. We’re generalists with a focus on vertical AI, robotics, climate, healthcare, and AI+bio. We mostly do software. We’ll back hardware when revenue is driven by software/data—e.g., deploying robots but monetizing the platform or data—not unit‑by‑unit hardware sales.
Jason Kirby: You’re an ML expert. What advice do AI founders ask for most—and how do you answer?
Prateek Joshi: Many start by making their AI offering more complex. My first advice: go the other way. If you can do it without AI, do that. AI is the nitro boost to an engine that already works—it’s not the engine. For fundraising, understand how equity is bought and sold. Generate real demand for your shares—sequential 1‑by‑1 outreach hurts pricing and process. Create parallel interest.
Jason Kirby: Differentiation is hard when everyone “adds AI.” How should founders truly differentiate?
Prateek Joshi: In early days it’s hard to be truly novel. When you can’t differentiate horizontally, go vertical. Don’t run from the big incumbent—embrace it: “We’re Snowflake‑like, but purpose‑built for this narrow customer set who prefer us 10× because our workflows are tailor‑made.” That gives you a beachhead of super‑fans. A few evergreen nuggets: follow your skill (not just passion); compounding takes time—don’t quit too early; learn to tell a story; master cold outbound; and remember—speed wins.
Jason Kirby: Where can people learn more or reach out?
Prateek Joshi: Visit pratikj.com for my writing, podcast, and books. Reach me at pratik@moxie.vc. If you’re a founder raising a seed round—or just want a sounding board—get in touch and mention this podcast. I read every cold email.
Jason Kirby: Prateek, thanks for joining and sharing your technical and business insights. Loved the frameworks and the candid stories.