Xiaoxiao Lei

Co-built an end-to-end AWS data pipeline of a financial instrument platform. The data pipeline ingests 30,000+ offering instruments/day from third-party vendor APIs, transforms data of 170+ attributes into the persistence layer, and serves downstream consumer applications. Built the platform's transformation and data-quality layer using PySpark on AWS Glue; strengthened data reliability and validity by normalizing vendor records into validated domain objects and enforcing a consistent data model; persisted canonical records to Aurora PostgreSQL and published enriched events through SNS.

Co-developed a Financial Data Enrichment Service (Spring Boot) on ECS for financial-data enrichment and serving, enabling users to preview, correct, and query edited records before downstream consumption; exposed GraphQL/REST APIs that delivered normalized views to consumer applications. Translated evolving business requirements into implementation-ready engineering specifications and used GitHub Copilot to accelerate implementation and test authoring, reducing feature delivery cycle time by app

Associate Software Engineer at J.P.

Morgan

Chase & Co. (2022-04 – Present)

• Co-built an end-to-end AWS data pipeline of a financial instrument platform. The data pipeline ingests 30,000+ offering instruments/day from third-party vendor APIs, transforms data of 170+ attributes into the persistence layer, and serves downstream consumer applications.
• Built the platform's transformation and data-quality layer using PySpark on AWS Glue; strengthened data reliability and validity by normalizing vendor records into validated domain objects and enforcing a consistent data model; persisted canonical records to Aurora PostgreSQL and published enriched events through SNS.
• Co-developed a Financial Data Enrichment Service (Spring Boot) on ECS for financial-data enrichment and serving, enabling users to preview, correct, and query edited records before downstream consumption; exposed GraphQL/REST APIs that delivered normalized views to consumer applications. Translated evolving business requirements into implementation-ready engineering specifications and used GitHub Copilot to accelerate implementation and test authoring, reducing feature delivery cycle time by approximately 70% across recent feature work.
• Developed a Document Broker Service (Spring Boot) that manages domain objects and supports document search, upload, and download backed by MongoDB; redesigned permission resolution as extensible flow for multi params access control; added MongoDB-backed audit logging to improve traceability.
• Strengthened production reliability and operational readiness by building an AWS Lambda utility that injected 30+ record types on demand for end-to-end testing, reducing debugging time by 80%; provided production support by monitoring and triaging alerts, tracing cross-service failures with Datadog and AWS CloudWatch, performing root-cause analysis, and coordinating prioritized remediation work through resolution.
• Own infrastructure delivery for an AI agent governance platform that validates architecture proposals against configurable rules. Build a repeatable, multi-environment AWS delivery foundation by provisioning cloud resources through Terraform IaC, deploying containerized services to ECS clusters, and manage CI/CD pipelines across environments with Jenkins, Jib, and Spinnaker.
• Built a Google ADK-based financial-instrument reconciliation agent system that compared data extracted from vendor emails against internal platform records. The system combined instrument-type routing, text extraction, and sequential/parallel comparison in a schema-guided multi-agent workflow using FastAPI, Azure OpenAI, and AWS Bedrock. In manually reviewed test emails, the prototype identified 95% of known discrepancies and was projected to save approximately 1.5 hours of daily manual review per banker.
• Delivered end-to-end development of a React-based admin portal for an internal fundraising platform, translating Figma designs and complex authorization rules into intuitive permission-management workflows; built reusable UI components within a monorepo and integrated the front end with GraphQL/REST APIs for managing read/write access to financial and entity records.

Master of Science in Computer Science and Engineering – State University of New York at Buffalo (2020-01 – 2022-02)

Bachelor of Arts in English Language and Literature – Beijing International Studies University (2010-09 – 2014-07)