Marshall Qiao
Senior Backend & Platform Engineer
A senior backend and platform engineer with more than a decade of experience in Java, distributed systems, and production delivery.
My experience spans fintech, social platforms, automotive IoT, e-commerce, and local services. I have worked deeply in merchant, catalog, and order domains, leading architecture, core implementation, and production delivery for high-concurrency, low-latency, highly available systems.
I have designed and evolved order systems supporting 1M+ orders per day, and built Kafka-based event-driven pipelines processing 300K+ events per second at peak and 1B+ events per day. My work also includes Spark and large-scale data processing, cloud delivery, observability, and failure recovery.
My current focus is AI development automation. I actively work with AI harnesses, RAG, and agent collaboration, integrating evidence traceability, review gates, safety boundaries, and verifiable delivery into real engineering workflows.
Work Experience
Meitou Inc.
Oct. 2025 — PresentSenior Software Engineer
Own the financial technology platform's backend end to end, spanning core business domains, data flows, and production infrastructure, from technology selection and architectural evolution through delivery and operations.
- Work across API contracts, serverless workflows, managed data, authentication, observability, and production operations.
- Deliver roadmap work across backend and web paths, from design and implementation through verification and release.
- Evolve event-driven AWS services and materialized read paths with explicit consistency, failure, and recovery boundaries.
- Optimized the end-to-end build, validation, and deployment workflow, reducing release time from approximately 60–80 minutes to under 40 minutes.
- Use evidence-based AI development practices with explicit review, safety, and delivery gates.
- TypeScript
- React
- GraphQL
- AWS
- Event-Driven Architecture
- Production Delivery
Soul
Oct. 2021 — Aug. 2023Staff Software Engineer
Led platform engineering for real-time risk control across feature computation, event processing, rule execution, and policy operations.
- Rebuilt the rule engine and redesigned Kafka/Spark streaming with Redis caching, moving minute-level P99 latency into the sub-minute range at the approved public scale bucket.
- Architected an end-to-end risk platform for rule and policy authoring, traffic replay, and A/B testing, materially reducing non-compliant content app-wide.
- Established system-design and code-review standards and coached a small group of engineers, improving team delivery efficiency.
- Spring Boot
- Kafka
- Spark
- Redis
- Elasticsearch
- Kubernetes
Tencent
Jul. 2021 — Oct. 2021Staff Software Engineer
Led core architecture and engineering delivery for an automotive IoT SaaS programme across a cross-functional engineering team.
- Owned critical code paths and established CI/CD, Agile, and project-delivery practices to improve quality and release velocity.
- Architected and delivered reusable cloud-service modules for an automotive SaaS platform used across multiple automakers.
- Java
- Spring Boot
- SaaS
- CI / CD
Alibaba Group
Jul. 2019 — Jul. 2021Senior Java Developer
Built open-platform and order-domain services for large-scale local-commerce and marketing programmes.
- Integrated cross-system data pipelines supporting a high volume of daily group-buying campaigns with reliable delivery.
- Unified fragmented PHP and Java front ends behind a single order-domain service and deployed active-active across multiple sites.
- Sustained a high-volume daily order business with stable low-tail latency while improving service availability.
- Java
- Spring Boot
- MyBatis
- Elasticsearch
- Python RPC
58 Daojia
Jul. 2017 — Jul. 2019Senior Java Developer
Owned merchant-platform backend delivery across operations, web, mobile, H5, service, and distributed-job systems supporting a large merchant base.
- Led sprint planning and merchant-side iterations, writing core code and introducing caching and rate limiting to improve stability.
- Modernised Spring Boot and Dubbo services, distributed-job deployment, observability with CAT and Graylog, and JaCoCo quality gates.
- Spring Boot
- Dubbo
- MyBatis
- Elasticsearch
- CAT
- Graylog
JD.com
Jul. 2016 — Jul. 2017Java Developer
Delivered full-stack e-commerce systems for large promotional campaigns and a crowd-creation platform.
- Built Java/Spring and Vue features that supported high-traffic campaign workflows.
- Contributed to architecture design and implemented core platform modules with scalability and business alignment in mind.
- Java
- Spring
- Vue
Hit Software
Aug. 2013 — Jul. 2016Full-Stack Developer
Delivered enterprise software from requirements analysis through implementation and client hand-off.
- Maintained strong client relationships through clear communication and translated business needs into maintainable software throughout the project lifecycle.
- Java
- PHP
- SQL
Selected Systems
A production platform, an evidence-first RAG system, and an independently operated learning product — viewed through my responsibility and key engineering decisions.
Production Financial Platform
Production · AnonymizedBackend and platform engineering for a private financial system, spanning API contracts, event-driven serverless paths, materialized reads, consistency controls, observability, and safer delivery.
Responsibility: Contributed across backend architecture, implementation, testing, and release operations.
Key decisions
- Keep request-time work bounded by moving suitable derivation into event-driven paths.
- Treat consistency changes and releases as observable state transitions with explicit recovery paths.
- TypeScript
- API Design
- AWS Serverless
- Event-Driven Systems
Evidence-first Developer RAG
Active development · Local shadowA local-first knowledge system that treats retrieved text as candidate evidence, not truth. Immutable content and version manifests back hybrid retrieval and an evidence graph, while grounded answers expose citations and unresolved gaps.
Responsibility: Designed and implemented independently in mt_rag; production cutover has not been performed.
Key decisions
- Separate immutable evidence from rebuildable vector, keyword, and graph projections.
- Let models rank and render within bounds; deterministic guards own chains, citations, and refusal.
- Python
- CAS + Manifest
- Hybrid Retrieval
- Neo4j
- Evaluation
PASS PTE
Founder · LiveA complete learning product with a Next.js frontend and Java/Spring backend, covering practice workflows, AI-assisted diagnostics, secure media, data operations, and reversible releases.
Responsibility: Founder and overall owner across product, engineering, infrastructure, releases, and operations.
Key decisions
- Wrap third-party AI with deterministic rules, validation, fallbacks, and reusable results.
- Use health-gated blue-green delivery so a product release remains observable and reversible.
- Next.js
- Spring Boot
- MySQL
- Redis
- Docker
For private systems, company details, internal identifiers, exact infrastructure counts, and business rules are intentionally generalized. The descriptions focus on my engineering scope and publicly shareable patterns.
Technical Skills
Frontend & Product
- React
- Next.js
- TypeScript
- HTML / CSS
Backend & Distributed Systems
- Java
- Python
- PHP
- Spring Boot
- Microservices
- GraphQL
- REST APIs
- Event-Driven Architecture
- DDD
Data & Messaging
- MySQL
- DynamoDB
- Redis
- Elasticsearch
- Kafka
- Spark
- Flink
Cloud & Delivery
- AWS
- Google Cloud
- DigitalOcean
- Alibaba Cloud
- Tencent Cloud
- Kubernetes
- Docker
- CI / CD
Applied AI & Retrieval
- RAG
- Hybrid Retrieval
- Qdrant
- Neo4j
- Evaluation
- AI Agents
Engineering Workflow
- Figma
- Jira
- Confluence
- Bitbucket
- Git
- Prometheus
- Grafana
Studies
Master of Engineering
2024 — 2025Electrical & Computer Engineering · University of Windsor
Bachelor of Engineering
2009 — 2013Information Security · Harbin Engineering University