ADITYA BAIRY
*** ********* ******, ********, *** York 13210
+1-315-***-**** ac8vng@r.postjobfree.com linkedin.com/in/adityabairy/ EDUCATION
Syracuse University, School of Computer Science May 2019 MS in Computer Engineering GPA: 3.7/4
Udacity School of Arti cial Intelligence Jan 2019 Nanodegree in Machine Learning Foundation and AI programming Bachelors of Science, Vidyavardhaka Engineering University May 2017 Department of Electronics & Communication Engineering GPA: 3.8/4 TECHNICAL SKILLS
Certi cation: IBM Data Science Professional Certi cation Computer Languages/Tools: Python, C, C++, R, Hadoop, HTML, CSS, MATLAB, Verilog Libraries/cloud: Keras, Scikitlearn, TensorFlow, AWS, SQL, Pandas, Pytorch, NLTK, Django Machine Learning Models: Logistic Regression, Random Forest, K Means Clustering Principal Component Analysis, Deep Neural Networks ACADEMIC PROJECTS
Design of Pacman game using deep reinforcement learning, CNN and OpenAI Gym Jan 2019
Exploited the OpenAI Gym MsPacman-v0 environment of size (210,160,3) to collect the data from the simulation
Estimated q-values using DQN, a deep CNN with three convolutional layers and two fully connected dense layers
Built a replay bu er called prioritized Replay to store past experiences (s, a, s, r) to use it to train Neural Net I-Screen: Eye gaze tracking system using convolutional neural network Oct 2018 - present
Collected data from GazeCapture, which contains facial image data and JSON les from over 1450 people
Successfully trained iTracker, a convolutional neural network using keras for the detection of user’s eyes.
Attained an error as low as 2.89 pixels away from the target, by evaluation of varied metrics
Performed ne-tuning and image augmentation on the pre-trained model to reduce the error considerably to 1.67 pixels away from the target co-ordinates
Multivariate Time Series Forecasting of Air Pollution at US embassy in Beijing using LSTM Oct 2018
Conducted exploratory analysis and visualization using Pandas and Seaborn on the Beijing PM2.5 dataset
Scaled, encoded and converted the Time Series data into a Supervised Learning data to feed the LSTM network
Evaluated by combining the forecast with the test dataset, inverting the scaling and achieving a test RMSE of 23.4 Election Interference Detection and Result Prediction Using Twitter and Facebook Oct 2018
Extracted and conducted sentiment analysis on tweets related to an election using extraction tools and NLTK library
Explored Bokeh Python library to create an application which lets user input a keyword and get the prediction result
Performed signi cant amount of feature engineering and developed a technique to detect any biased Ad bots RELEVANT COURSES AND PROJECTS
Courses Projects
Introduction to AI implementation Poppy Humanoid
Machine Learning Foundation Boston housing price prediction Social Media Data Mining Sentiment analysis of Twitter and Yelp Object Oriented Design Source code to webpage conversion ACADEMIC EXPERIENCE
Graduate Assistant for Advanced Computer Architecture under Dr. Ehat Ercanli Aug 2018
Gained in-depth knowledge through research, helped curate assignments, research papers, exams and graded them