Design NFT Lending Protocol - System Design | WinJob

Problem Statement: NFT-Backed Collateral Loans

NFTfi / BendDAO / Blur Lending-class protocol where borrowers lock ERC-721/1155 collateral in escrow, lenders supply WETH/USDC liquidity pools, and loans settle via on-chain loan NFTs with time-bounded liquidation auctions. This understanding section (1/30) focuses on problem for NFT-backed lending.

Mechanism

Borrowers post NFT collateral; lenders fund pooled WETH/USDC at utilization-driven APY
Each loan is a transferable position NFT with principal, rate, expiry, and liquidation threshold
Floor-price oracles plus TWAP bands gate max LTV; liquidators bid in Dutch auctions on default

Lending invariants

On-chain escrowed collateral and loan principal are authoritative; off-chain floor feeds are projections that must carry stalenessSeconds and confidence metadata. Health factor recomputes on every material floor move.

Failure modes

Thin-collection floor spoofing during borrow (sec-01)
Double-liquidation if auction coordinator lacks idempotency on loanId
Lender bank-run when utilization hits 100% and withdraw queue stalls

Interview checkpoint (sec-01)

When NFTfi or BendDAO interviewers probe problem, cite $800M peak TVL, 8k concurrent loans, and HF < 1 triggers Dutch auction—not a generic DeFi diagram.

Staff+ talking point

Separate borrower UX latency from lender solvency: pools can stay solvent while previews slow down—never hide bad debt behind cached APY billboards.

javaOne Dark Pro

1 public record LoanHealth1(java.math.BigInteger floorWei, java.math.BigInteger debtWei) {
2   public boolean underwater() { return floorWei.multiply(java.math.BigInteger.valueOf(80)).compareTo(debtWei.multiply(java.math.BigInteger.valueOf(100))) < 0; }
3 }

pythonOne Dark Pro

1 def health_factor_1(floor_eth: float, debt_eth: float, threshold: float = 0.8) -> float:
2     return (floor_eth * threshold) / debt_eth if debt_eth else float('inf')

typescriptOne Dark Pro

1 export function healthFactor1(floorWei: bigint, debtWei: bigint, thresholdBps = 8000n): number {
2   if (debtWei === 0n) return Number.POSITIVE_INFINITY;
3   return Number((floorWei * thresholdBps) / debtWei) / 10000;
4 }

Why interviewers care

NFT Lending Protocol interviews reward crisp scope, explicit trade-offs, and failure stories—not generic microservice diagrams.

Interview checkpoint

Name one failure story for Problem Statement: NFT-Backed Collateral Loans that proves you understand real outages, not happy-path diagrams.

Key Highlights

•Borrowers post NFT collateral; lenders fund pooled WETH/USDC at utilization-driven APY
•Each loan is a transferable position NFT with principal, rate, expiry, and liquidation threshold
•Floor-price oracles plus TWAP bands gate max LTV; liquidators bid in Dutch auctions on default

1	public record LoanHealth1(java.math.BigInteger floorWei, java.math.BigInteger debtWei) {
2	public boolean underwater() { return floorWei.multiply(java.math.BigInteger.valueOf(80)).compareTo(debtWei.multiply(java.math.BigInteger.valueOf(100))) < 0; }
3	}

1	def health_factor_1(floor_eth: float, debt_eth: float, threshold: float = 0.8) -> float:
2	return (floor_eth * threshold) / debt_eth if debt_eth else float('inf')

1	export function healthFactor1(floorWei: bigint, debtWei: bigint, thresholdBps = 8000n): number {
2	if (debtWei === 0n) return Number.POSITIVE_INFINITY;
3	return Number((floorWei * thresholdBps) / debtWei) / 10000;
4	}