About

Harsh Maheshwari is a highly skilled and experienced professional in the fields of Quantitative Finance, Data Science, Deep Learning, and Consulting. He is a graduate of the Indian Institute of Technology, Bombay, and has also completed an exchange semester at DTU - Technical University of Denmark. He is currently working as a Business Intelligence Intern at Axis Bank, where he has built a micro-service for predicting customer income using nominatim API, scikit-learn, docker, and kubernetes. Harsh has also engineered a new feature based on customer address using K nearest neighbor algorithm and geocoding services, and obtained an accuracy of over 90% using a XGboost regression model based on customer history and transaction data. He has developed a web interface using Django framework and containerized it in a Docker container for easier deployment. He has reduced daily up time by using Kubernetes orchestration for automated managing and scheduling of services. Harsh has worked as a Computer Vision Lead at Innovation Cell, IIT Bombay, where he designed and developed an autonomous fixed-wing for search and rescue missions customized for the Barcelona Smart Drone Challenge. He has integrated a ROS-enabled camera system for detection of characters from a height of 50 meters on a flying drone. He has also developed a multi-core zoom in character recognition model based on Char74 dataset using OpenCV & Keras, and obtained an accuracy of 94% using transfer learning on MobileNetV2 architecture on custom aerial data set. He has also worked as a Research in Artificial Intelligence Intern at NUS Computing, where he developed a Temporal Attention model for node classification and link prediction tasks in different temporal graphs. He has understood the dependency of graphs on time using dynamic models like Node2Vec and Temporal Node Embedding, and found inefficiencies in the above models and improved them to get AUC of 86% for College Messages dataset. Harsh is skilled in Python, Matlab, R, bash, SQL, HiveQL, HTML, CSS, Matplotlib, Numpy, Pandas, SciPy, OpenCV, Ggplot2, dplyr, PIL, Pytorch, Keras, Tensorflow, Scikit Learn, NLTK, spaCy, FastText, SysID ToolBox, LSTMs, CNNs, GANs, Naive Bayes, Decision trees, Regression models, Random Forests, K-means, PCA, SVM, Apache Spark, SparkML, MySQL, Sqlite,