Backend engineer with a passion for building production systems that scale.
I'm a backend engineer with 3+ years building production systems in fintech and data platforms. I currently ship the core services behind Syfe's fund products across Hong Kong and Singapore — owning product surfaces end-to-end, from requirement to production-ready system.
My work centres on distributed systems, event-driven architecture and clean, maintainable design: Kotlin/Spring Boot microservices, Kafka pipelines, multi-currency financial flows and the observability that keeps them healthy. If you're building something ambitious on the backend, I'd love to hear about it.
Experience
- Own the Income Max HK product backend end-to-end (Kotlin/Spring Boot microservices) — portfolio-type creation, custom security allocation, multi-currency FX handling (HKD/USD), ex-date dividend accrual, iFile generation, payout reconciliation and reinvestment. Income Max HK USD has become the fastest-growing portfolio by AUM in the HK market.
- Architected the SRS (Singapore) price-ingestion pipeline from scratch with an event-driven PSPL file ingestion job, Slack alerting and a reconciliation service — eliminating manual price uploads and ops intervention.
- Led backend for Cash+ Fixed Portfolio Operations covering the full lifecycle: maturity, subscription, spread accounting and a configurable bank/client-rate approval workflow.
- Designed and shipped a Structured Change Logs framework (fluent builder API, AOP @TrackChanges aspect, Strategy-pattern sink) — collapsing 6-line audit call sites to 1 across 50+ sites (85% less boilerplate) with full backward compatibility.
- Built AI-powered engineering workflows with Claude Skills — an automated code-review agent and an on-call debugging assistant integrating Datadog, Metabase and Sentry — cutting production incident triage time by 50%.
Education
Aug 2019May 2023
B.E. Electronics & Instrumentation Control
Thapar Institute of Engineering and TechnologyProjects
- Built a context layer for AI coding tools (Claude Code, Cursor, Windsurf) that captures architectural decisions, conventions and gotchas into a structured knowledge graph, then compiles relevant subsets into the native context files each tool reads automatically (CLAUDE.md, .cursor/rules) — no manual copy-paste across tools.
- A local daemon (Node + Fastify + SQLite) watches files, compiles context and syncs to a Postgres-backed cloud API for multi-machine consistency, fronted by a sub-5ms Go CLI client.
- Shipped as a pnpm + Turborepo monorepo spanning the agent, CLI, a VS Code/Cursor/Windsurf extension, cloud API and a Next.js docs site over a shared typed core.
- Designed a single security platform for AI agents with two products on one account, dashboard and versioned event schema: Sentinel (prompt-injection prevention) and Blackbox (action recording & rewind).
- Sentinel guards each agent run with a honeypot, a mitmproxy-based firewall and an output scanner, returning a SAFE / COMPROMISED verdict per run.
- Blackbox records every action an agent takes as reversible inverses, enabling rewind(to_step=N) to roll execution back to any prior step. Go engine + Next.js dashboard + Python SDK in a pnpm monorepo.