Cellular Physical Layer Firmware Engineer

Aspiring to develop technologies for Electric vehicles at affordable prices and better parameters like average speed, acceleration and range.• Configured Bootloader on the microcontroller over a secure (CSEc module) communication protocol obtained through OTA (Over-the-Air) updates.• Implemented the Battery Management System and communication system of the vehicle using the Controller Area Network (CAN) communication protocol.• Tested GPS/GSM module for implementation in the… Show more Aspiring to develop technologies for Electric vehicles at affordable prices and better parameters like average speed, acceleration and range.• Configured Bootloader on the microcontroller over a secure (CSEc module) communication protocol obtained through OTA (Over-the-Air) updates.• Implemented the Battery Management System and communication system of the vehicle using the Controller Area Network (CAN) communication protocol.• Tested GPS/GSM module for implementation in the vehicle; Implemented REST API and MQTT API to publish data online (through POST, GET requests). Show less

Project: COVID-19 detection from chest X-rays images using various deep learning methods and machine learning classification techniques.• Analyzed the behaviour of multiple layer CNN architectures with classification parameters such as accuracy, F1-score, ROC AUC area and Kappa scores.• Used machine learning classification techniques on top of features extracted by the various pre-trained models to train data of 6553 chest X-Ray images.• Critically examined the change of performance… Show more Project: COVID-19 detection from chest X-rays images using various deep learning methods and machine learning classification techniques.• Analyzed the behaviour of multiple layer CNN architectures with classification parameters such as accuracy, F1-score, ROC AUC area and Kappa scores.• Used machine learning classification techniques on top of features extracted by the various pre-trained models to train data of 6553 chest X-Ray images.• Critically examined the change of performance brought about by transfer learning with different pre-trained models of Keras of which VGG 16 achieved the best results to get an accuracy of 94%; 6-CNN, MobileNetV2 and DenseNet121 models achieved the highest COVID-19 classification accuracy of 97%. Show less

Project: To deevelop a realistic application algorithm to monitor the sleep quality of an individual unobtrusively using an off-the-shelf smartphone.• Separated noise frames using specific variance features to build an ambient noise model to calculate the threshold of features for detecting sleep events.• Developed a lightweight decision tree-based algorithm using extracted acoustic features to classify sleep events with an EDA accuracy of more than 93%.• Detected sleep efficiency… Show more Project: To deevelop a realistic application algorithm to monitor the sleep quality of an individual unobtrusively using an off-the-shelf smartphone.• Separated noise frames using specific variance features to build an ambient noise model to calculate the threshold of features for detecting sleep events.• Developed a lightweight decision tree-based algorithm using extracted acoustic features to classify sleep events with an EDA accuracy of more than 93%.• Detected sleep efficiency using activity score based on Actigraphy and PSQI based sleep quality estimation to achieve an FDA accuracy of more than 90%. Show less

• Analyzed 36 million customers booking data using 32 features to segment five different types of customers that OYO serves using unsupervised learning.• Developed dashboards for tracking customer feedback and property audits on a real-time basis using Metabase to reduce UH feedbacks from 13% to 11%.• Ideated and developed a code to schedule customized automated operations reports as mail, daily to the ground team for daily targets they need to hit.