BZ
Binliu Zhang's Doodles

My Projects

Case studies from food delivery, mobility, and fintech platforms where I led automation, data, and experimentation initiatives.

Operational Excellence

Help Experience Orchestration Layer

Event-driven automation powering JET’s help chatbot. The orchestration layer coordinates services, state, and playbooks so order issues resolve without manual intervention.

Impact:

  • Automated resolution flows for millions of annual customer contacts, reducing manual interventions by 60%.
  • Improved stability and MTTR with resilient retries, stateful orchestration, and structured observability.

What I did:

  • Owned architecture, API contracts, and rollout strategy across product, operations, and support teams.
  • Partnered with ops to codify playbooks into reusable workflows, guardrails, and experimentable policies.
JavaSpring BootKafkaAWS SQSRedis
Experience Overview
Workflow diagram for help experience orchestration
Platform Engineering

JETMS Operational Platform

Multi-tenant workspace that unifies 20+ internal services for customer support agents. CQRS routing, shared data models, and fine-grained permissions keep teams focused on resolution outcomes.

Impact:

  • Reduced issue-resolution SLA by ~33% by consolidating fragmented tooling into a single experience.
  • Enabled market-specific workflows with tenant-aware routing and flexible schema evolution.

What I did:

  • Design service decomposition, and front-end framework-agnostic architecture.
  • Leverage CQRS patterns and governance for integrating partner teams.
JavaTypeScriptMicro FrontendsMySQL
Experience Overview
Screenshot of JETMS operational console
Geo Platforms

MapYourCity Crowdsourcing Platform

Crowdsourced geospatial data engine fueling Grab’s mobility and delivery products. Coordinates contributor tasks, validates submissions, and feeds downstream routing services.

Impact:

  • Scaled to 300K tasks and 400K+ geospatial data points that improved map freshness and trip accuracy.
  • Cut geospatial storage costs by 70% via data simplification and tiered retention policies.

What I did:

  • Designed task lifecycle, task modeling, and auditing capabilities across data, ops, and product teams.
  • Implemented high-performance services in Go, PostGIS, and Kafka with proactive monitoring and alerting.
GoPostgreSQLPostGISKafkaElasticSearch
Live Demo Experience Overview
Map and geospatial data pipeline illustration

Other Projects

Additional initiatives that strengthened observability, experimentation, and developer velocity.

Experiment dashboard mockup

Darwin Experimentation Platform

Re-architected experimentation APIs and SDKs at Ant Group to support 1,000+ annual experiments and drive a 2% conversion lift.

JavaSDKExperimentation
View Project →
Street-view map tiles

KartaView Street-View Service

High-performance PostGIS service applying the Douglas–Peucker algorithm to shrink geospatial datasets by 70% for faster queries.

GoPostGISMicroservices
View Project →