Ashutosh Bansal

Engineered a computer vision-based data science model for real-time analysis of cab drivers' selfies, utilising image processing algorithms to discern uniform adherence with precision, optimising operational workflows. Developed a baseline model from scratch to predict OLA's electric scooter's remaining range in real-time (using telemetrics data of 40k+ scooters on road). The model was tested on 2k scooters for Move OS 3 beta version, resulting in 86 % accuracy. Contributed to the enhancement of Ola cabs' pricing model, specifically focusing on airport routes, by incorporating demand and supply dynamics based on microeconomic principles. Leveraged advanced data analysis techniques to quantify demand patterns, supply availability, and market conditions. Implemented a comprehensive digitisation solution for cab driver onboarding documents using OCR technology. Leveraged OCR algorithms to extract text and data from documents such as Aadhar, PAN, permits, and more. And integrated them into the document processing pipeline to automate data extraction and minimize manual effort. Built an advanced tire pressure model leveraging machine learning (without any sensor data), utilising classification algorithms and alerting user for low tyre pressure.

Developed period tracker app analyzing PMS symptoms, predicting period characteristics using Time Series modeling and LSTM neural networks for Ovulation/Fertility Window prediction.