Working in clickstream team by building tools to catch regressions to metrics and monitoring tools with alarms to cut tickets to source tickets for changes in data.

Worked on marketing data ingestion starting from 3p to ingestion of internal data. Built data quality validation pipeline which caught lot of data issues. Integrated with anomaly detection to catch issues uncaught by heuristic approaches. Built pipelines for transformation and data models to store and process the data. Built infrastructure to make data ready for model ingestion that is feature engineered.

Built dashboard tool to automatically load the model output data and display results stakeholders in addition to data health.

Built AC classification model to predict the type based on speed and use on PyTorch using random forest classifier with accuracy of 88%. This was crucial as the whole backend system required the system type to process the data which will show up on the customers application.

Built regression model to predict temperatures used to evaluate the AC working mode. This is very critical to determining the AC health.

Built time series forecasting model to determine the cycle time which is necessary to classify the health of the system.

• Designing/Developing/Testing/Optimising software to analyse HVAC systems health.

• Researched and analysed efficient ways of storing large amount of data (RDS/DynamoDB/Kinesis) and design decisions(Throttling) involved.

• Created and investigated data driven visualizations to build an end-to-end system

• Developed experimental and analytic plans for data modeling and feature engineering.

• Applied various machine learning techniques, and analysed and understood the key parameters that affect the performance.

• CI/CD involves statistical analysis, optimizations, regression modeling and forecasting, time series analysis of the data collected from IOT devices.

Working on VBS 2021-fast video segmentation and annotation for International Conference on Multimedia Modelling(MMM2021) using techniques of computer vision, machine learning, deep learning, convolutional neural networks, natural language processing, text analysis and retrieval.

Contact: Dr Seon Kim