Launched real-time ASR model-build multi-threading GRPC microservice system.

Took charge of both offline time transcription API and real-time model build release MVP to partner team in Q1 2024.

Realtime video stream automatic data pipeline to transmit video for model training

Product daily active user growth starts with 10-20 users to more than 200 per day.

Implemented Client interface to connect ASR ML services and handle audio traffic

built-up machine learning microservice in the task queue system

Optimized table structure and migrated table from SQL to NoSQL database for reading optimization

Reduced nearly 20% cloud service cost by setting up VPC endpoint, S3 glacier archived level class

The revenue increases 12% monthly and the customer sign-up rate increases 19% monthly after

Developed ETL pipeline with AWS SNS/SQS to fetch and store promotion data in the database.

Designed RESTful APIs to generate shared URLs so they can be shared with other customers

Built database table schema and define primary keys to optimize high-volume customer read queries

Created script and configuration to generate server health metrics automatically and reduce 20% the operation effect.

Setup the web application in cloud computing Server- Microsoft Azure

Use cosine similarity to automatically calculate partial credits for exam questions and enhance grading consistency

Design front-end in responsive pages with HTML5, CSS, jQuery/JavaScript, Bootstrap

Visualize the data model on abstract student learning behavior

Continue supporting application develop in future ASU cs courses