Design DeFi Insurance - System Design | WinJob

Problem Statement: On-Chain Mutual Insurance

Nexus Mutual pioneered discretionary mutual cover where members stake ETH into a pooled capital model and vote on claim payouts after smart-contract incidents. This section anchors problem statement: on-chain mutual insurance during the understanding phase of a Nexus/InsurAce-class design.

Mechanism

buyCover transfers premium to pool; activeCover increases until MCR blocks new sales.
Risk engine recomputes product weights hourly from exploit frequency and severity priors.
Claims attach IPFS evidence hashes; assessors stake reputation tokens and vote during a time-boxed window before escrow releases payout.

Solvency invariants

On-chain activeCover + claimReserves ≤ bondedCapital × MCRfactor is evaluated atomically on every buyCover. Off-chain dashboards may lag one block but purchase simulation must mirror contract revert reasons.

Failure modes

oracle spike triggering false parametric payout on thin stablecoin pool
liquidity crunch when stakers rush to withdraw after large approved claim

Interview checkpoint (sec-01)

State 72h assessment window and 67% quorum for discretionary payouts.

Staff+ talking point

Contrast discretionary vs parametric honestly — hybrid products need two audit surfaces.

javaOne Dark Pro

1 public final class McrGuard {
2   public boolean canSellCover(long activeCover, long capital, int mcrPct) {
3     return activeCover * 100L <= capital * mcrPct;
4   }
5 }

pythonOne Dark Pro

1 ACTIVE_COVER_USD = 800_000_000
2 BONDED_CAPITAL_USD = 1_120_000_000
3 MCR_RATIO = BONDED_CAPITAL_USD / ACTIVE_COVER_USD

typescriptOne Dark Pro

1 export interface CoverQuote {
2   productId: string;
3   coverAmount: bigint;
4   durationDays: number;
5   premiumWei: bigint;
6 }

Why interviewers care

DeFi Insurance interviews reward crisp scope, explicit trade-offs, and failure stories—not generic microservice diagrams.

Interview checkpoint

Name one failure story for Problem Statement: On-Chain Mutual Insurance that proves you understand real outages, not happy-path diagrams.

Key Highlights

•Cover capacity is bounded by **Minimum Capital Requirement (MCR)** — active cover amount m
•Off-chain services index `CoverPurchased`, `ClaimSubmitted`, `ClaimVoted` events into Post
•72h assessment with 67% quorum

1	public final class McrGuard {
2	public boolean canSellCover(long activeCover, long capital, int mcrPct) {
3	return activeCover * 100L <= capital * mcrPct;
4	}
5	}

1	ACTIVE_COVER_USD = 800_000_000
2	BONDED_CAPITAL_USD = 1_120_000_000
3	MCR_RATIO = BONDED_CAPITAL_USD / ACTIVE_COVER_USD

1	export interface CoverQuote {
2	productId: string;
3	coverAmount: bigint;
4	durationDays: number;
5	premiumWei: bigint;
6	}