I build AI-native, production-grade backend systems — trading infrastructure, real-time data at scale, on-chain and enterprise platforms. Currently SDE III at American Express; ex–founding engineer at Filament Finance, ex–Goldman Sachs and BlackRock.
I’ve built the parts that can’t go down: a perps matching engine with order books, TP/SL and a comb-pool model at Filament, and a decentralized perpetuals exchange across 20+ commodity, crypto and forex markets at Carry. Comfortable with order lifecycle, funding, liquidations and sub-second APIs.
Streaming pipelines that hold up under load: 10M+ events/day in Rust over NATS JetStream with exactly-once delivery and checkpoint recovery, landing in ClickHouse at 490M+ rows. I care about backpressure, replay and cost, not just throughput.
My default architecture at Goldman Sachs, BlackRock and ZestMoney: Java / Spring Boot microservices on Kafka, event sourcing on Cassandra, big-data pipelines on Spark/Hadoop, and AWS-native async flows with SQS/SNS, Lambda and Step Functions.
Smart contracts that move real value: on-chain collateral & PnL at Filament, Solidity trading/vault/funding/oracle contracts with ERC-4626 vaults and Pyth at Carry, and a Quorum tokenization network for warehouse receipts at Whrrl.
I ship things a business depends on: a dual-database DB2/PostgreSQL migration framework at Amex hitting 99.9% consistency, and a farmer-lending platform at Whrrl with SAP integration, video KYC, and a 50+ bank maker/checker loan workflow that I own end to end.
I ship faster by working with Claude Code, Cursor and Codex as part of my workflow — using agents and skills to move from spec to production. The judgement, architecture and review stay mine; the leverage is the point.
Architected an enterprise dual-database DB2/PostgreSQL framework with independent transaction boundaries and a 124-parameter transformation pipeline, plus production monitoring reaching a 99.9% data-consistency benchmark. Built a blockchain lending protocol for crypto assets at the AMEX hackathon.
Built a decentralized perpetual trading platform for BTC and SEI from the ground up — order book, comb-pool model and a Spring Boot / Kafka / Redis / MongoDB backend sustaining 60 orders/second, with on-chain collateral and PnL settlement.
Built trade-flow and post-trade surveillance controls across the front-to-back architecture using Java / Spring Boot / Kafka microservices, with big-data pipelines on Spark and Hadoop.
Built event-driven loan creation, settlement and closure microservices on Java, Spring Boot and AWS (SQS/SNS, Lambda, Step Functions, S3) with MySQL, plus risk-team APIs and a Cucumber BDD suite.
Built ETF order-fulfilment backends for primary and secondary markets using event-sourcing microservices on Java, Spring Boot, Kafka and Cassandra, with resiliency and retry for completeness.
Real-time streaming pipeline in Rust ingesting 10M+ events/day across 20+ streams on NATS JetStream, exactly-once with checkpoint recovery into ClickHouse (490M+ position rows). Trader analytics, leaderboards and wallet classification over REST + WebSocket.
Decentralized perpetuals exchange with Solidity contracts (trading engine, vault, funding, oracle) and a Go backend across 20+ commodity, crypto and forex markets. 40+ endpoint REST API, TradingView charts, dynamic funding, automated liquidations, ERC-4626 vaults and Pyth.
Warehouse-receipt tokenization on a Quorum private chain letting 35,000+ farmers borrow against stored produce. Node.js / Express + Angular with real-time SAP integration, video KYC, S3 docs and a 50+ bank maker/checker loan workflow. I own the AWS footprint and reconciliation.
Decentralized perpetual trading for BTC and SEI with a comb-pool model. High-performance order book on Spring Boot / Kafka / Redis / MongoDB sustaining 60 orders/second, with on-chain collateral and PnL settlement.
I treat Claude Code, Cursor and Codex as part of the toolchain, not a novelty. Agents and skills let me go from a clear spec to reviewed, production code far faster than typing it all by hand.
What doesn’t change: the architecture, the trade-offs and the review are mine. I’m accountable for what ships, and I read every line before it does.
Rust · Go · Java · JavaScript · Python · SQL
Spring Boot · Node.js · Express · Angular · Actix
Kafka · NATS JetStream · SQS/SNS · Lambda · Flink
ClickHouse · PostgreSQL · MySQL · MongoDB · Redis · Cassandra
Solidity · EVM · Quorum · Pyth · ERC-4626
Docker · Kubernetes · AWS · Nginx · Prometheus · Grafana