Senior Consultant (Individual Contributor)

Aurix is a Non-profit initiative to create a global AI community.I created the Education vertical of Aurix, providing free resources, classes and explanations on the concepts of AI.

My master's degree consisted of 50% coursework and 50% research. I broadly worked on "Air Quality Estimation"• Experimenting with portable air quality sensors (Jan 2019 - Apr 2019)Measured the exposure of PM2.5 on construction workers using air quality sensors. Plotted the data acquired over time and conducted calibration tests with 5 sensors. • Creating a live air quality dashboard (May 2019 - Jul 2019)Deployed 5 air quality sensors at different locations on campus.… Show more My master's degree consisted of 50% coursework and 50% research. I broadly worked on "Air Quality Estimation"• Experimenting with portable air quality sensors (Jan 2019 - Apr 2019)Measured the exposure of PM2.5 on construction workers using air quality sensors. Plotted the data acquired over time and conducted calibration tests with 5 sensors. • Creating a live air quality dashboard (May 2019 - Jul 2019)Deployed 5 air quality sensors at different locations on campus. Created a live dashboard using JavaScript, python(flask), and leaflet.js that displayed the amount of PM2.5 at those locations and updated the values automatically after an interval. This study was useful in measuring high polluting activities and their time in those 5 locations.• Estimating ground air quality using satellite data (Aug 2019 - Dec 2019)As air quality sensors are expensive and provide less spatial coverage, we used satellite-derived aerosol optical depth (AOD) and meteorological fields collocated in space and time to estimate surface-level PM2.5 using a regression model.• Curating a dataset for India and using Machine Learning (Jan 2020 - Jul 2020) Gathered PM, AOD, and meteorological data for India through various sources and collocated them in time and space. Used Ridge regression, random forest, time-series cross-validation for our analysis like seasonal variation in pollution. This study helped us in identifying the challenges for data curation and collocation for India. Show less

• Project: 'Botnet detection using deep convolutional neural networks'• Generated a dataset containing normal and botnet data in pcap format, converted the pcap files into images and fed them into different CNN models.• Compared and analyzed the performance of VGG16, ResNet, AlexNet and GoogLeNet on classification of botnet and normal data.