Python Software Engineer

Adarsh Verma

Linkedin: www.linkedin.com/in/verma-adarsh Email: ******@*****.***

SUMMARY

• 3 years of experience as a Java/Python Developer in the IT field which includes Machine learning, Textual analytics, Big data.

• Hands-on experience in writing a python scripts for various kind of natural language processing, machine learning, and computer vision problems. Great understanding of Amazon’s EC2 setup and maintenance activity.

• Proficient in Python and Java, implemented several machine learning algorithms like Support Vector Machine (SVM), Linear regression, Random Forest, K-Means, Naïve Bayes, Perceptron, ANN with Backpropagations, Association Rule mining. EDUCATION

Master of Science in Computer Science November 2019 Rutgers University, New Brunswick, NJ, USA GPA: 3.23/4.0 Bachelor of Technology in Computer Science May 2012 J.S.S. Academy of Technical Education, Noida, India GPA: 3.2/4.0 TECHNICAL SKILLS

Programming Language: Java, Python, C++

Framework /Libraries: Spark, Amazon EC2, Apache Mahout, Apache OpenNLP, JSoup, Scikit-Learn, Apache POI Version Control/Tools: Git, Tortoise SVN, Putty, WinSCP, SOAPUI Database: SQL Server, Oracle 11g, MySQL

Web Technologies: P5.js, D3.js, AJAX, REST, HTML, CSS, Spring, JSP Cloud Computing Services: Amazon Web Services, Microsoft Azure PROFESSIONAL EXPERIENCE

Travelex India, India March 2013 – November 2016

Software Engineer (Java, Python)

• Sales Prediction Analysis – Sales patterns were analyzed on a three-month period for each store separately, and projections were predicted based on these. Stock management was advised based on these projections as well.

• Product Classification - This project was to develop a sentiment analysis mechanism on Scala and Spark. As an input, we received customer reviews for various products and services of the client. Analyzed these reviews and find polarity of the review based on the computed score using open NLP packages.

• SQL Server to Hadoop Migration - Data migration utility to migrate existing system into Hadoop (Big Data) from SQL Server system. Developed history creation process of data warehouse application and perform aggregation operations on Hadoop Big Data platform. PROJECTS

Implementing range queries in object-oriented databases July 2011 – May 2012

• Implemented range queries in OODBs by generating has coded signatures of the attributes of the database and storing them in a signature de-clustering tree

• Observed that the query retrieval time was enhanced by 43% under cases when the signature tree was weighing more than 50% of its capacity

Predicting emotions by speech analysis January 2017 – May 2017

• Implemented a prediction model in Python to detect emotions in speech with accuracy of approximately 74%.

• Designed categorical generative adversarial network (GAN) as our prediction model.

• Extracted features from the raw speech files by converting them into spectrograms using fast Fourier transform. Generating music using character generating RNN September 2017 – December 2017

• Generated new music in the form of ABC (musical) notations by training a Long short-term memory (LSTM) RNN in Python.

• The model learns the pattern of the text of the ABC notation, learning the musical comprehension of the genre as well, and then generates new ABC notation of music for each genre. Measuring level of toxicity in Wikipedia comments January 2018 – May 2018

• Attempted the Kaggle challenge hosted by Google which tried to measure the toxicity of comments and tried to categorize the comment as insults, threats, bullying etc.

• Implemented Gated Recurrent unit RNN model for this, as it performed better than other models like LSTM etc. Achieved approx. 71% precision when the threshold of toxicity was set to 0.7 probability. Implementing text ranking of articles in java September 2018 – December 2018

• Implemented Text Ranking for articles which could be used to search articles by keywords.

• Achieved this by assigning ratings to keywords of articles and making connections between the article based on their keyword rating.

• Designed GUI where users can input keywords and receive articles that have highest rating of keywords, and suggestions can be made as per similarities between article keywords.

Contact this candidate