Machine Assistant

Samip Thakkar

480-***-**** ****************@*****.***

linkedin.com/in/samip-thakkar-1310

github.com/samip-thakkar/

EDUCATION

Arizona State University, Masters in Computer Science (August 2019 – May 2021) 3.71 / 4 Gujarat Technological University, Bachelors in Computer Science (Graduated May 2019) 9.01 / 10 EXPERIENCE

Application Development Intern – Machine Learning (Upcoming) Jan 2021 – May 2021 ADP (Automatic Data Processing) Pasadena, California Research and Technical Assistant Sep 2019 – Present Arizona State University Tempe, Arizona

Built an application which auto grades the paper review using Natural Language Processing concepts. It grades on basis of originality, entropy, length of response and sentiment analysis.

Developed an application that extracts results of students from Slack using API and calculate and display the results visualization and mail the students their result.

Working on US Navy resilient research assisting in software development and data analysis. Machine Learning Intern May – July ‘18

Sculptsoft Ahmedabad, India

Worked on Research team for data analysis and research for a project on student loans.

Learnt about various Machine Learning algorithms. Worked on a Time-series data for forecasting and achieved 95% confidence with ARIMA model.

TECHNICAL SKILLS

Programming Languages: Python, SQL, JavaScript, Java, C++, HTML/CSS.

Frameworks: Flask, Django, Pytorch, TensorFlow, Keras, AWS, GCP, Azure

Tools and Databases: Git, Hadoop, Spark, Tableau, MongoDB, MySQL PROJECTS

Gender and Biometrics recognition using for Image Data

Computing image similarities from dataset of unlabelled 11,000 hand images.

Used dimensionality reduction, feature extraction, classification, clustering, feedback system, Locality sensitive hashing and page ranking.

Auto-scaling using AWS and Raspberry PI

Live Object Detection using Raspberry pi and AWS services like S3, EC2 and SQS.

Reduced the latency using Auto-scaling, multi-processing, and parallel computing. Auto grading text using Natural Language Understanding

Auto grading the student’s multiple text responses based on Originality of work, sentiment analysis, word count and constructiveness in the answers. Web scraping was done for data extraction. Fraud Detection using Graph Database

Built a fraud detection model for transactions using graph database and graphQL by Neo4j. Web application was built by Flask and deployed on Google Cloud Platform.

Increased model AUC-ROC score by 5-10% using sampling and graph features. Medical Corpus Entity Linking using BERT

Built a Named-Entity Normalizer using BERT to link the similar words and similar concepts together. The accuracy was 83% more than 7% by SOTA using Elmo and Bi-LSTM. Machine Translation using Locality Sensitive Hashing

Built a naïve Machine Translation for converting an English word to French using linear translation of embeddings. Implemented document search using Locality Sensitive Hashing.

Lot more projects on github.com/samip-thakkar/

Contact this candidate