I come across as an inquisitive person, trying to uncover the mysteries behind data by utilizing various techniques and models of data science. My internships, projects, various online courses and academic courses have guided me in solving such data oriented problems. Along with this, I love to understand the underlying mathematics and workings of the machine learning models I come across, which in turn has led me to develop such models from scratch. Some of these include the naive bayes model, logistic regression with l2 normalization, decision trees, markov decision model, k-nearest neighbor with cross validation.
Apart from the data domain, I have always had an interest in Android as well as Web development. To this end, I have completed Certification courses ( in both), an academic course and an internship in Android, as well as worked on various projects.
-
Experience
1. Suven Consultants and Technology Pvt Ltd (SCTPL)
• Researched the effect on the climate of various cities based on their proximity to the ocean. The OpenWeather API was used to collect weather data dynamically for each city.
• Tested the hypothesis involving the digit dataset of sklearn library.
• Extracted association rules from a grocery dataset using the Apriori algorithm.
2. Suven Consultants and Technology Pvt Ltd (SCTPL)
• Developed an android application for a distributor to solve the issue of low productivity of their delivery agents by providing online logging of sales data and recording of the GPS location as well as
the time of delivery of products. Data was stored and retrieved from a Realtime Google Firebase.
• This application improved the productivity of the agents which in turn increased the profit margin of the distributor by 8% over the period of a month.
3. Indian Railways
• Developed a predictive crowd estimation tool using the XGBoost Model in Python.
• Displayed the results with graphical visualizations on a website.
• The model helped the railways redirect trains to crowded stations.
• The project was presented and received approval from the Senior Engineer at Central Railways.
-
Projects
1. Handwriting Matching using Neural Networks
• Developed a deep learning based handwriting matching system in Python.
• Used the opencv library to perform image processing and break the sentences in the image to words.
• Built an algorithm working in an erode-average fashion to further segment the words into characters.
• Used the keras library to build a Convolutional Neural Network classifier which took the characters as input and provided the percentage likeness as output.
• Presented the project to a domain expert and stood first in the university.
2. Gopher Hunting
• Developed an AI based game playing android application between two players in Java.
• The game consists of 100 tiles out of which one houses a gopher. The two players must guess a tile
and the game continues until either of the two guess the tile consisting of the gopher.
• Each player has a different heuristic which helps the player making a guess as the game progresses.
3. NotiForex
• Developed an android application to display the conversion rates of different currencies and provide
updates through notifications based on a user specified interval.
• The application also displays a notification when a major change in the selected currency occurs.
• The application uses a client-server service interaction to fetch data from an API in the background.
• The data was obtained through freeforexapi in the JSON format and then parsed by the application.