Archan Dutta
** ***** *******, *** ****, CA, ****4
ac6ime@r.postjobfree.com, Ph: 408-***-****
LinkedIn: https://www.linkedin.com/in/archan-dutta
EDUCATION
University of Southern California, Los Angeles, CA May 2018
Master of Science, Computer Science (GPA: 3.4/4.0)
Devi Ahilya University, Indore, India June 2016
Bachelor of Engineering, Information Technology (GPA: 3.7/4.0)
TECHNICAL SKILLS
Specializations: Data Science, Machine Learning, NLP, Algorithms, Databases
Programming Languages: Python, R, SQL, PHP, Node.js, JavaScript, Java, C#, C++
Technologies: AWS (Amazon Web Services) S3, EMR, EC2, Google Firebase, IBM Watson,
Hadoop/MapReduce, Spark, MongoDB, REST API, GitHub, JIRA
WORK EXPERIENCE
Konviv, Berkeley, CA, (Software Engineer Intern – Data Science) (Jun 2017 - Aug 2017)
Built a full-stack application for financial assistance that allows users to make smarter spending decisions. Completed 50% of the production level goal and worked closely with the business and product teams.
Achieved an accuracy of 94% using Support Vector Machines to classify each transaction into one of the six spending categories. Generated insights about the overall spending habits of the customer.
Technologies/ Tools: Node.js, JavaScript, Firebase, Plaid API
Indian Institute of Technology, India, (Data Science Intern) (May 2015 - Aug 2015)
Implemented N-gram language model on Brown Corpus (1 million unique words). Improved the model performance using Laplace Smoothing and Turing Smoothing.
Created a Search Engine based on PageRank algorithm. Enhanced the Search Engine by implementing Auto-complete feature using the N-gram Language model.
Improved the existing system by 300% measured on the basis of Perplexity of the N-gram model.
Technologies/ Tools: Java, PHP, Python, Apache Solr, Apache Lucene
Arcadia Riptides Swim Club, Los Angeles, (Software Engineer Intern) (Sep 2017 - Dec 2017)
Developed a website that facilitates registration for participants in swimming events.
Improved the business workflow by 92.3% based on the benefit analysis. Expected 95% Return on Investment by the year 2020.
Technologies/ Tools: MySQL, JavaScript, PHP, Apache PDFbox
MACHINE LEARNING PROJECTS
Project: Statistical Analysis of Capital Bikeshare Data (May 2018 – Jun 2018)
Trained a Logistic Regression to predict member type with a F-measure of 0.91 on an imbalanced dataset containing 4 million data points.
Performed Data Wrangling, Visualization, Feature Engineering and Multivariate analysis.
Technologies/Tools: Python (Pandas, Matplotlib, Scikit-learn), Amazon Elastic MapReduce (EMR)
Project: Scyther (Data Scientist and Game Designer) (Jan 2018 – May 2018)
Designed a FPS game where the player interacts with a NLP based artificial intelligence system using a speech recognition interface. Primarily responsible for defining the game-play, script and implementing the artificial intelligence system, camera behavior, map and event triggers.
Technologies/ Tools: Unity 3D, C#, IBM Watson (Assistant, Speech to Text)
Project: Unsupervised Emotion Detection of Hindi Songs (Algorithm Engineer) (Feb 2017 - Apr 2017)
Obtained F-measure of 0.86 on emotion detection of Hindi language songs using Spectral Clustering and NLP techniques. Conducted a survey that obtained 81% inter-rater agreement measured by Kappa statistic.
Technologies/ Tools: Python (NumPy, Beautiful Soup, NumPy)