Sign in

Medical Data

Tempe, AZ
February 25, 2020

Contact this candidate



480-***-**** linkedin/dhruvpatel Github/dhruvpateldp96 YouTube/dpdhruvpatel1996 EDUCATION

Master of Science in Computer Science Expect May 2020 Arizona State University, Tempe, AZ,USA CGPA 3.67/4 Bachelor of Technology in Computer Science May 2018 Nirma University, India CGPA 7.14/10


Languages: Python, Java, JavaScript, C++

Technologies: AWS, HTML5, CSS3, MapReduce, Hadoop, Spark, SQL, Android, ROS, Tableau, D3.js, MATLAB Database: MySQL, NoSQL-MongoDB

Tools: Eclipse, Pycharm, Git and Github, Anaconda, Docker, numpy, pandas, TensorFlow, Pytorch, Jupyter Platform: Windows, Linux SDLC: Agile, Waterfall


Indian Institute for Plasma Research Intern Jan 2018 - April 2018

Developed an intelligent airborne vehicle for exploring and mapping the internals of the Nuclear fusion reactor. Similar to GRASP lab’s (UPenn lab) RAPID project .

Implemented 3D SLAM algorithm using 2D lidar. Created fake point cloud data by fusing laser scans and sensor data.

Leveraged the use of octree, for robust mapping of the environment and navigating complex 3D environments. link to -3d Hector Slam implementation

Keywords: Docker,Robot Operating System, Hololens, objective-C, C#, Linux, Open Source, OpenCV, PCL. Yochan Lab Graduate Research Assistant Jan 2019 - May 2019

Worked under the guidance of Dr. Subbarao Khambhampati (AAAI President 2016).

Developed an Autonomous Reconnaissance Vehicle for Urban Search And Rescue (USAR) scenarios that can collaborate with Human operators to optimize exploration.


Autodidactic model for Medical Image Segmentation and Classi cation (based on MICCAI-2019) Apr 2019

We trained the model in a self-supervised manner using unlabeled image datasets. It still matched the performance of the traditional supervised transfer learning models trained on ImageNet.

We achieved an improvement of 3% in 3D semantic segmentation by this approach. Deep Q-Learning and Pacman Nov 2018

Created an intelligent Pacman agent using raw pixel data as input. (Implemented Deep Q Network from scratch)

Used TensorFlow with Python to design the CNN and trained the algorithms against smart ghosts. Medical diagnosis using Patient’s Medical Records Apr 2018

Created an NLP model for topic modeling and knowledge extraction from a large corpus of Bio-medical documents using LDA and Doc2Vec models.

Used Elasticsearch for fetching out the relevant documents before applying the algorithms. Geospatial Data Clustering for Tra c Management Nov 2019

Developed software for classifying dense tra c areas (for tra c management), using Apache Spark and spatial queries for geospatial GPS data.

Reinforcement Learning based Robot Playing Basketball (OpenAI paper) Apr 2019

Overcame sparse reward problem in reinforcement learning, by implementing the HER algorithm (based on OpenAI’s NIPS-2017 submission). (e.g.Robot playing basketball {has rare reward signals) 3D Object Pose Detection and Classi cation for Autonomous cars (Lyft Competiton) Oct 2019

Merged data from multiple sensors mounted on the car and performed point cloud segmentation using pointnet (CVPR 2017). Thereby calculating the pose of various obstacles. Multimedia Storage, Indexing and Retrieval software Nov 2019

Compressed Multimedia data using algorithms like SVD, PCA, LDA and stored it in the MongoDB database.

Implemented graph partitioning algorithms like KD-trees for e ective storage and PPR and LSH for retrieval of data. ACHIEVEMENTS

Semi-Finalist in DRUSE-2017 (DRDO Robotics and Unmanned Systems Exposition). DRDO is India’s Defence Research and Development Organisation.

Contact this candidate