Jul 2019 – Jul 2021 · 2 yrs · Gurugram, Haryana, India

• Quantitative Research Development team @ Limestone.
• Signal generation and trading strategies execution infrastructure in C++.
• Market data research and post-trade analysis tooling in Python.

May 2018 – Jul 2018 · 2 mos · Tokyo, Japan

• Worked on feature embedding indexing for similar image search optimization given a query image.
• Reduced search time from 720 ms to 8 ms on a dataset of 30 million images at 88% 1 recall at 5.
• Implemented network distillation in TensorFlow to train lightweight MobileNetV2 model from InceptionV3 model.

May 2017 – Jul 2017 · 2 mos · Greater Hyderabad Area

• Developed an XGBoost model to predict clients' subscription renewal chances with 90% accuracy.
• Integrated ML model interpretation tooling to help the sales team identify opportunities to reduce churn.
• Developed an LSTM-based time series prediction model of the users’ online activities using Keras.

Feb 2016 – Mar 2017 · 1 yr 1 mo · IIT Kharagpur

• Team AGV participated in IGVC 2016 (Intelligent Ground Vehicle Competition) at Oakland University, Michigan, USA. We are also finalists competing for 1 million USD in the Mahindra Rise Prize - Driverless Car Challenge.
• Designed an heuristic algorithm for lane-navigation based on Maxwell-Boltzmann curve using RANSAC after processing camera feed and LIDAR data.
• Successfully implemented and tested two different path planners, one was implemented using the MRPT library, and the other using the move_base framework of ROS.