Piyush Kurkure

linkedin.com/in/piyushkurkure adm36d@r.postjobfree.com 619-***-****

http://www.github.com/Piyush-Kurkure San Diego, CA Education

San Diego State University, San Diego – CA, USA Expected May 2021 Master’s in Computer Science

Related Coursework: Applied Computer Vision, Big Data Tools and Techniques, Full Stack Development, Modern Web Frameworks, Combinatorial Algorithms, Spatial Databases. University of Mumbai, Mumbai – MH, India August 2015 - May 2019 Bachelor’s in Computer Science

Related Coursework: Data Structures, Designing Algorithms, Object-Oriented Programming, Computer Networks. Work Experience

Tata Institute of Fundamental Research, Mumbai, India — ML Research Intern Apr 2018 – May 2019

• Built an automated system using a DNN algorithm (YOLO network) that would be capable of detecting real-time ships across the borders and secured industrial areas with impressive fps of 45 and speed of 22ms/image.

• Developed a surveillance system that can detect a maximum of 20 ships and marine vessels in pre-captured videos as well as real-time video streaming to increase portability.

• Implemented the project on Qt application so that it can be used across the organization by all engineers and operators.

• Build a model that is extremely fast, more than 100x times the Fast R-CNN and was trained on 330K images. Siemens, Mumbai, India —Software Developer Intern Dec 2017 – Jan 2018

• Developed a troubleshooting manual/quick reference application that will be used by people working in the industry to post, answer, delete and update the reviews for any problems faced with the pieces of machinery in the industry.

• This tool helped workers present in 24 different departments to resolve their problems without concerning the managers of their departments. This automated monitoring reduced the manual work by 60%.

• Implemented live chat option between users by socket.io, reducing average response time by 100%. Avics Consultancy Services, Pune, India — Software Developer Intern Aug 2017 – Sept 2018

• Developed a digit recognizer model which would correctly predict the labels for numerical digit’s image passed to it.

• The model was tested on 28000 samples to classify image into 10 different classes.

• Using the deep learning concepts, increased the accuracy up to 95% and collaborated on deploying the model on system. Project Work

Quora Question Pairs

• The project was aimed to predict whether the pair of questions have the same meaning out of provided pairs of two questions. The dataset consists of around 404,300 question pairs for identification.

• Data analysis and visualizations are performed to see how questions are asked by people on the Quora platform.

• Technologies: Keras, NLTK, sklearn, plotly, LR model, LSTM model TED Data Analysis and recommender

• The project was aimed to analyze the TED talks data having information about all talks including the number of views, number of comments, descriptions, speakers and titles.

• Build a recommendation system to suggest the Ted talk to any user based on their previous watched talks.

• Technologies: Pandas, Seaborn, PySpark, sklearn, Vaex, NLTK Titanic Survival Prediction

• Project aimed to develop a predictive model that tells us what sorts of people were more likely to survive during the sinking of the Titanic. Concepts of Feature Engineering and transformation are used to increase the precision up to 85%.

• Technologies: Pandas, sklearn, TensorFlow, Keras, decision trees, random forest classifiers. Skills and Tools

Programming Languages Python, JavaScript, C++, Java, HTML, CSS. Databases Relational (MySQL, MS SQL), Non-relational (MongoDB). Tools & Platforms Git, AWS, Docker, Django, Django-rest, Flask, REST, Excel. Data Science Hadoop, Spark, Dask, Vaex, NumPy, Pandas, TensorFlow, Keras, Pytorch, Sklearn. Other Skills User Experience Designing, Product Management, Research, Communication Skills, HCI.

Contact this candidate