Lead Software Engineer
at Unbanked
I was responsible for leading the integration of multiple debit and prepaid card services issued with white label partners who wanted to launch a card that could use their token for payments. Accomplishments include plug-in system implementation, team management, and enhancements to internal system architectures.
Spending crypto in real life
At Unbanked, we built a debit card that let people spend cryptocurrency, like Bitcoin or Solana, at normal stores. Users could swipe the card at Starbucks, and the money came from their crypto wallet, not from a bank account.
I worked there for two years as a Lead Software Engineer. The job was hard, because we had to connect the blockchain world with the traditional banking system. Over 30,000 people used our cards.
Two weeks per partner
When I joined, the company was growing, but it had a problem. We worked with crypto companies like StormX and Zebec to create branded debit cards for them. This is called a white label program. We built the card, and the partner put its own name on it.
Adding each new partner took two full weeks of custom code. We had to write new code for each cryptocurrency and set up everything by hand. With ten new partners, the work would take five months.
I built a White Label Plugin System to fix this. Partners no longer needed custom code. They entered their cryptocurrency details, like the token address and the blockchain, and the system connected it to our banking system automatically.
Adding a partner went from two weeks to just minutes. That change helped us grow much faster. The new partner integrations brought $1.2 million in revenue, and the partner work drove 2 million in revenue in a single year.
Adding Solana support
Next, we worked with Zebec, a big company in the Solana ecosystem. They wanted their own branded card.
Adding Solana was harder than Ethereum. Solana is very fast, but it is also more complex. Our system was not built for how Solana works.
I built a custom system with @solana/web3.js that checked the Solana blockchain for new deposits all the time. It had to be fast, so users could use their card right after sending cryptocurrency to their wallet. If we missed a check even once, someone might have their card declined at a store.
We made it work. Thousands of users could now spend USDC and SOL with our cards, which connected Solana to the Visa payment network. I also rebuilt parts of our internal architecture so the system could support Solana and other blockchains. After that work, it could integrate thousands of assets.
Working directly with partners
The plugin system removed the coding bottleneck. The human side still needed an owner.
I held the technical relationship with around 10 partner programs, including StormX, Zebec, Uberstate, United Coin and United Token. Their tokens lived on different blockchains. StormX, Uberstate and CrownSterling were on Ethereum. Zebec brought USDC on Solana. United Token and United Coin were on Stellar.
I ran their onboarding, integration, requirements and troubleshooting calls. Then I led the backend changes each program needed. I also worked with our card vendors, CashPass and Tribe, and coordinated fixes across their systems and ours.
Partners should not have to wait for a call to keep moving. So I wrote the public partner documentation at docs.getblockcard.com. It covered webhooks, custodial integration, KYC (Know Your Customer, the identity checks banks must do), rewards and testing endpoints.
I designed the webhook platform they integrated with. A webhook is a message our system sends to a partner’s system when something happens, like a new deposit. It came with replay tools for missed deliveries. I also built communication subsystems with sockets, which keep a connection open between two systems. And I built simulation endpoints, so partners could test card and deposit flows without moving real money.
Leading from inside the team
I led a team of up to 5 engineers, but not from a distance. For each partner requirement, I put together a small team from whoever was capable and available. I ran the dailies and owned the scope of the domain. I implemented the core features myself, and I delegated the blockchain and white label subtasks.
This horizontal style worked, because everyone could see me shipping the same code they were.
Systems that protect money
When you work with people’s money, bugs are serious. A mistake could mean someone loses money.
We needed very reliable systems. I led the work to rebuild our background processing with BullMQ and Redis. We moved transaction processing, fraud checks and email notifications into queues. A queue holds jobs until a worker is ready, so those queues could handle thousands of operations at once without slowing down. The improved caching, jobs and queues cut memory use by 80%. They also cut the load from open connections on the caching server by 60%.
We also needed a way to test the card system without spending thousands of dollars. I built a Virtual Card Terminal that could simulate every situation, like declined cards and timeouts. We tested thousands of scenarios every time we deployed new code. Our error rate dropped from 10% to almost zero.
Compliance ran through the same systems. I automated OFAC sanction suspensions, false positive handling and KYC status flows across all the card programs. OFAC is the US office that keeps the list of sanctioned people and countries. When regulators require you to act on a sanctions match, “we’ll do it manually” is not an answer.
How it ended
Unbanked made the Inc. 5000 list of the fastest growing companies. We processed millions of dollars in transactions. But startups are risky. Our technology worked well, and still the company ran out of funding and had to close.
It was sad to see the company shut down. But what we built is one of my proudest achievements. We showed that cryptocurrency could work as real money, not only as something to invest in.