Jan 2022 – Jul 2022 · 6 mos · Trollhättan, Vastra Gotaland County, Sweden

• Simulated extreme scenarios to generate synthetic dataset with annotations for improving object detection in Autonomous Vehicles using the CARLA simulator (based on Unreal Engine).
• Set up pipeline to evaluate uncertainty of object detector predictions using Probabilistic Object Detection.
• Evaluated impact of factors that affect prediction uncertainty like change in weather/lighting conditions.

May 2021 – Jul 2021 · 2 mos · Bengaluru, Karnataka, India

• Designed neural networks using CNN and LSTM layers in Keras for enabling On-Device classification of text messages to provide smart action/reply buttons to smartphone users.
• Deployed the same for Samsung’s flagship phones using Tensorflow Lite to achieve low latency and reduction in resource consumption by 86%.

May 2020 – Jun 2020 · 1 mo · Chennai, Tamil Nadu, India

• Developed and designed a neural network using LSTMs trained on Amazon Reviews Dataset (300,000 reviews) that predicts customer rating (1-5) of unlabelled product reviews at 79% accuracy.
• Deployed the model to predict customer ratings on unlabelled reviews scraped from sites like Product Hunt to provide insights on current market sentiment of products chosen by user.

Jul 2019 – Jul 2019 · 0 mo · Jaipur, Rajasthan, India

• Selected among top 40 candidates across India from multiple IITs, NITs and BITS
• Improvement of android-based application for in-house data collection streamlining.