Responsible for designing and engineering the sensor based smell level classification system used by the built odour meter

Designed and developed the data acquisition strategy to collect artificial data; collecting balanced data of good, normal and bad smell signatures of more than 10,000 data points; from within Xavor’s washroom.

Developed a custom smell classification pipeline; with an accuracy of 94%; using a combination of supervised learning (Xgboost) and unsupervised learning (K-Means Clustering) algorithms, deployed as a Flask service docker container on Azure

Machine Learning, Data Analysis, REST/RESTful APIs, AWS Glue, Flask, Python, Docker, Modular APIs, Azure, ETL, Pandas, Numpy, Git