jaydenng

Designed and implemented a scalable and fault-tolerant AI GuardRails batch processing pipeline using FastAPI as the user interface, S3 for object storage, Snowflake for data warehousing, Autosys for scheduling, Jenkins for CICD, Bedrock as LLM-judge and AWS servers for infrastructure. Improved processing speed of large gigabytes files from 2 weeks to 3 days, enhancing efficiency by ~70%. Developed a full stack AI governance platform using NextJS and Python FastAPI that automates the company's AI lifecycle management process. The platform enables users to submit their proposed use cases and track their approval progress, while providing management with effective data analytics capabilities through React dashboarding and audit logging functionality. Designed and implemented an automated Extract, Transform, Load (ETL) data transformation platform that onboards new data sources from databases and file based systems into Snowflake, and performs change data capture (CDC) from landing to staging and to the consumption zone for MI reporting. Improved data onboarding speed from an average of 1 week to 1 day, enhancing efficiency by ~50%.

Designed and implemented Micro-Frontend Architecture for the Crypto and Equities Trading app using React, Typescript and Redux. Executed features including Module Federation, Analytics Tagging, Error Handling, Data Table Filtering and Sorting, routing, Backend API integration, Real-time streaming, Design library integration, and optimization. Implemented web APIs including authentication workflow and real time equities' parameters streaming using Springboot; Implemented step-wise new trader assessment workflow using Camunda, Kafka and Microservices; Performed container configuration, deployment and testing in dev to UAT env. Performed Test-Driven Development using Jest and Pact, and utilized Cypress for end-to-end integration testing with Backend; Utilized JUnit for Springboot unit testing.

Collaborated with clients to integrate intricate business logic and complex financial calculations into clients' new Hadoop infrastructure; Implemented complex data transformations using SQL to develop an automated target table view within Hive and Impala for ease of accessibility by data scientists. Developed and optimized data pipelines to migrate clients' corporate data from diverse sources including CSVs, PDFs and Parquet files, enterprise CRMs like SalesForce, and relational databases like MSSQL and Postgres into the Hadoop ecosystem via a combination of Shell, Python and SQL scripting.

Managed a team of 6 junior engineers, driving task allocation, process monitoring improvements, and mentoring to enhance overall team efficiency in delivering high-quality technical solutions. Spearheaded the data initiative of building an analytics dashboard using Python for analysis and PowerBI for MI, targeting critical equipment to improve performance monitoring and failure prediction using Seeq and Power BI, which brought about potential cost benefits of USD 100K per year. Initiated the complementary use of data analysis techniques, including statistical inferential testings and correlational analytics, into the team's daily monitoring plan to improve anomalies tracking capabilities, which ultimately led to the preventive catch of a critical malfunction saving up to USD 200K.