Skip to main content

PRODUCT-MINDED DEVELOPER

Development changed.
So did I.

Fullstack developer with product thinkingBuilding products, not just writing code

The old model (PM designs, developer codes) no longer works. AI accelerated development dramatically, but created a new problem: teams need people who understand both what to build and how.

I combine both skills. 8+ years in development, 5 in web3 across EVM and Solana. Worked in startups where there's no luxury of separating roles. Took product management courses. Won Ethereum hackathons.

Latest project: basepaint.market. I noticed BasePaint collectors had nowhere to sell complete year collections (365 NFTs). OpenSea and Blur don't support atomic bundle trades. I designed the solution, wrote the smart contract, backend, frontend, conducted an audit, launched to mainnet. Solo.

View WorkGet in Touch

SELECTED WORK

Products I've Built

2025 · LIVE ON BASE MAINNET

BasePaint Market

Visit Site

The first specialized marketplace for trading complete BasePaint year collections on Base L2. Enables buying and selling complete sets (365 NFTs) in a single atomic transaction. Something impossible on OpenSea or Blur.

3,767

Automated Tests

9.0/10

Audit Score

365

NFTs per Bundle

0

Critical Issues

SolidityNext.js 14Node.jsPostgreSQLRainbowKitBase L2

2026 · BASEPAINT YEAR 3 HACKATHON WINNER, FOR COLLECTORS

Missing Days

Visit SiteGitHubWinners Announcement

A BasePaint profile for any address. Paste a wallet and see which daily canvases it owns and, the useful part, which days are missing from each set, which of those gaps are scarce, and what it painted. No wallet connect, no keys, no database: every number comes from BasePaint’s public indexer.

Judges: “Clean, simple and instantly useful to visualize and complete a collection.” Given away CC0 so it can fold back into BasePaint itself.

1st

FOR COLLECTORS

32

Projects Entered

0

Backend or Keys

CC0

Public Good

Next.js 16React 19viemVercelBasePaint IndexerCC0

2026 · CHROME EXTENSION

Curfew

Install ExtensionGitHub

Daily time budgets for the distraction sites you choose, enforced at the network level. Relaxing a rule costs a solved puzzle instead of a click, so the decision stays deliberate. Local-only: no accounts, no analytics, and no network layer of its own.

0

Network Requests

Per-site

Permissions

200

Unit Tests

MIT

Open Source

JavaScriptManifest V3Chrome APIsDeclarative Net RequestNo Tracking
More work

WHY PRODUCT-MINDED?

Questions I Get Asked

"AI can code now. Why hire a developer?"

AI removed the cost of typing. It did not remove the cost of deciding. It doesn't know what to build, what to leave out, or which tradeoff is worth taking, and it will happily produce a working version of the wrong thing.

So the job moved up a level: hold the goal, cut the scope, pick the pattern (escrow vs approval-based listings), know what to verify and why, and keep pushing after the third dead end. That is the part I do, with AI as the fastest tool I have ever had.

"Why not hire a PM and developer separately?"

In a startup, there's no time for the PM → Designer → Developer → QA chain. When one person sees the user's problem and solves it themselves, the cycle shrinks from weeks to days.

Example from BasePaint Market:

Collectors didn't know the real value of their sets on the secondary market. Instead of creating a ticket and waiting, I built an Analytics Dashboard in a day: daily floor price on OpenSea, day rarity scores, number of complete set holders, and missing NFTs that can't be purchased. The feature strengthened the core product and gave users the data they needed.

"What does 'product thinking' actually mean?"

It's the ability to see a user's problem and find a solution through the product, not just write code to a spec.

Example:

On OpenSea/Blur you can't sell 365 NFTs in one transaction. Why? No atomic bundle transfers. Solution: a smart contract with approval-based listings (not escrow: less gas, safer). Result: the first specialized marketplace for BasePaint collections.