Leo Polovets

Who they are

Leo Polovets is General Partner at Humba Ventures — a Caltech CS grad who was LinkedIn's second engineering hire and co-founded Susa Ventures (early backer of Robinhood and Flexport) before launching Humba to bet on nuclear reactors, drone swarms, and surgical robots.

Person

Leo took his Caltech CS degree straight into some of the most consequential early-stage engineering roles of the 2000s — he was LinkedIn's second engineering hire, building core products before the company was a household name, then moved to Google on payments fraud detection, and later to Factual on big data infrastructure. That operator track record — early, technical, infrastructure-level — is what makes his VC identity coherent rather than coincidental. In 2012 he co-founded Susa Ventures, a seed-stage software fund whose portfolio grew to include Robinhood and Flexport. He launched Humba Ventures in 2022 as a sister fund to Susa, this time aimed squarely at deep tech: nuclear microreactors, drone swarms, surgical robots, genomic platforms, and AI-enabled defense. He writes prolifically — a blog, Medium, and a Substack — publishing frameworks on startup evaluation, the 'bundles of risks' de-risking approach, fundraising mechanics, and deep tech investing. He's appeared on Software Engineering Daily, The Full Ratchet, Heavybit's Venture Confidential, and a string of other podcasts under the label 'The Coding VC.' The through-line across three decades is operator-turned-investor who reads risk at the infrastructure layer before anyone else does.

Company

Humba Ventures closed its oversubscribed second fund — Humba II — at $40 million in December 2024, the most recent notable milestone for the firm. The fund targets deep tech and defense tech at the seed stage, writing checks into companies building nuclear microreactors, drone swarms, surgical robots, and AI-enabled defense solutions. Recent portfolio activity includes a May 2024 seed investment in Hexium, which is building a decentralized network of compact isotope separation units for fusion fuel, and a 2025 seed participation in Adaptyx Biosciences. The firm claims a 60%+ seed-to-Series-A conversion rate and aggregate portfolio valuations exceeding $10 billion. Humba operates out of San Francisco, sharing space with Susa Ventures in the Mission District, with co-founder Anna-Sofia Lesiv — who came from 8VC, Founders Fund, and Bridgewater — leading the firm's 'How It's Built' thought leadership.

Market

Defense tech venture investment hit record highs in 2024, with startups in the sector raising nearly $3 billion, creating strong tailwinds for Humba's thesis. Humba competes in this space against larger platforms — notably Andreessen Horowitz's American Dynamism initiative and Founders Fund — though Humba differentiates as a specialized seed fund with a deep technical operator as lead, rather than a generalist growth platform. The broader hard-tech seed market rewards firms that can evaluate technical risk early; Humba's positioning leans on that engineering credibility.

Network

Anna-Sofia Lesiv co-founded Humba Ventures with Leo and comes from 8VC, Founders Fund, and Bridgewater — she runs the firm's thought leadership vertical. Beyond the firm, Leo's network traces back to the LinkedIn founding team and the early Susa LP base, which carried over into Humba II. No additional named edges are available from current data.

Anna-Sofia Lesiv· Co-founder, Humba Ventures (formerly 8VC, Founders Fund, Bridgewater)

How they likely show up

Engineering roles at LinkedIn (second hire), Google, and Factual before moving to VC → evaluates companies at the infrastructure and technical-risk layer first, not the market-size layer.
Founded two VC firms (Susa Ventures in 2012, Humba Ventures in 2022) rather than joining existing platforms → high agency, builds from scratch rather than inheriting someone else's thesis.
Publishes structured frameworks (bundles of risks, startup evaluation, fundraising mechanics) on blog, Medium, and Substack → thinks in systems and likes to make his mental models explicit and teachable.
Podcast appearances span operator-focused shows (Software Engineering Daily) and investor-focused shows (Full Ratchet, Heavybit) → comfortable in both technical and financial conversations, code-switches fluently.
Humba II was oversubscribed at $40 million with LPs carrying over from Susa → long-term relationship builder; trust compounds across fund cycles rather than being rebuilt each time.
Possibly — based in Dallas per one signal, while the firm is San Francisco-based → may operate with a distributed or remote-first approach to portfolio engagement.

Conversation tips

→ Reference a specific framework from his blog or Substack — the 'bundles of risks' model is a good anchor — he'll know immediately whether you've read him or are faking it.
→ Ask about the operator-to-investor transition specifically: he's spoken about it on the Visible.vc podcast and it's clearly a theme he finds worth unpacking.
→ Engage on the technical diligence process for hard tech — how does a software-background investor evaluate a nuclear microreactor? He has a worked-out answer and will enjoy giving it.
→ Don't treat Susa and Humba as interchangeable — they have distinct theses (software vs. deep tech) and he separated them deliberately; show you understand the distinction.
→ He responds well to precise technical framing; vague market-size pitches will land flat with someone who was debugging payments fraud at Google.

Toolbox

Openers

Open on the Hexium investment — Humba backed a company building compact isotope separation units for fusion fuel in May 2024, which is about as far from traditional software VC as you can get. It's a direct signal of where his conviction sits right now.
Reference his 'bundles of risks' framework from his blog — he's built a public, structured methodology for how to evaluate and sequence de-risking in startups, and leading with it shows you've done real homework.
Mention the Humba II close in December 2024 — oversubscribed at $40 million — and ask what changed in LP appetite for hard tech between Fund I and Fund II. It's a concrete recent event with a real story behind it.

Discovery questions

Your background is software infrastructure — when you're evaluating something like a nuclear microreactor or a drone swarm company, where do you start the technical diligence and what does 'de-risked enough for seed' look like?
Susa and Humba have distinct mandates — software vs. deep tech — but share LPs and office space. How do you manage the thesis discipline so the two funds don't blur over time?
Defense tech hit nearly $3 billion in venture investment in 2024. Does a rising tide in that category make seed-stage differentiation harder, or does it actually help Humba because it validates the exits?

Avoid

Don't conflate deep tech with software — he explicitly separated the two into different funds, and treating nuclear, robotics, and defense investments as just another software bet will signal you haven't understood his thesis.

Make it yours