CASE STUDY / 02

Dexlyn Perps

A perps venue is not a chart with leverage. It is a live risk engine, an execution promise, and a trust surface.

Dexlyn Perps cover
ROLEProduct lead, risk modeling & execution
MODELGMX-v1-style -> CLOB research
METRIC18s -> <3s order placement
LEVERAGEup to 150x

Context

A perps product is a live risk engine wrapped in a UI. Every pair needs decisions: position caps, skew handling, funding behavior, open-interest limits, liquidation mechanics, collateral, execution speed.

What I did

Spearheaded research, documentation, and pair-level risk modeling. Wrote the scripts defining skew factors, skew scale, open-interest limits, funding rates, and funding velocity per pair. Led collateral partnerships with Supra-native stablecoin partners. Led the data-aggregation team for cost-effective lifetime trading data. Drove the latency effort that cut order placement from 18 seconds to under 3 - across transaction paths, backend processing, indexing freshness, and frontend state. Led research on adding orderbook functionality to the GMX-style base, and currently lead research into a Lighter-inspired CLOB perps architecture: matching, margin, liquidation keepers, maker incentives, and the on-chain/off-chain split.

How it works

Funding rates are incentives disguised as math - each pair's parameters tell traders what the venue rewards. The 18->3s change wasn't polish: below a few seconds a trader believes the venue is real; above it they assume it's broken.

Impact

Moved Dexlyn beyond spot into a trader-centric, risk-sensitive surface - and produced the research base for the next market-structure generation.

NEXT PROJECTDexlyn Aggregator & Bridge