Machine Learning Engineer

Sri Harsha Satya Kumar Konuru

**** ******* *****, *********, ******** 21227, USA

ac270p@r.postjobfree.com +1-732-***-****

EDUCATION

University of Maryland Baltimore County, Baltimore, Maryland, USA

Master of Science (M.S.) in Computer Engineering Aug 2015 – May 2017

Cumulative GPA: 3.2 /4.0

GITAM University, Hyderabad, Telangana, India

Bachelor of Technology (B. Tech) in Electronics and Communication Jul 2011 – June 2015

Cumulative GPA: 3.6 / 4.0

SKILLS

Coding Skills: C, C++, Python, R, HTML, CSS, Bootstrap, SQL, JSON, MPI, CUDA (Parallel and distributed computing), Java, Hadoop, Tensor Flow, Torch

Software Skills: Microsoft Visual Studio, Git (Git and Mercury repositories) Matlab, Xilinx, Android.

Operating Systems: Windows 2003/7/8/10, Ubuntu, Mac OS

Computer Skills: Microsoft Office, Visio, Adobe Suite

WORK EXPERIENCE

The Lobo Lab, UMBC

Machine Learning Research Associate, Biology Department July 2017 - Present

Worked on developing machine learning model (evolutionary algorithm) on gene equations.

Improvised the computing using GPU’s by implementing parallel threads on CUDA cores.

Developed the application using QT libraries to represent the gene hierarchy.

Technology: CUDA with object oriented C++, Microsoft Visual Studio, Evolutionary algorithm.

Energy Efficient High Performance Computing Lab, UMBC

Machine Learning Research Assistant, Computer Science Department May 2016- July 2017

Worked on deploying machine learning models on biomedical data (EEG, Heart rate)

Developing websites in EEHPC for lab and maintaining them. Responsible for coding, innovative design and layout of the website.

Create website layout/user interface by using Bootstrap and using JavaScript, PHP, HTML and CSS.

Designed and implemented a GUI application using QT creator for demonstration of a project.

RESEARCH PROJECTS

An EEG Artifact Identification Embedded Hardware using ICA and Multi-Instance Learning

A novel software-hardware system that uses a weak supervisory signal to indicate that some noise is occurring.

The EEG data is decomposed into independent components using ICA, and these components form bags that are labeled and classified by a multi- instance learning algorithm

Technology: Python 3 using numpy and scikit-learn packages, C++, CUDA for parallel implementation, R, Matlab, Machine Learning (Independent Component Analysis, Principal Component Analysis, Multi-Instance learning, SVM)

Raspberry pi Hadoop Cluster based Data processing

Created a Raspberry Pi based cluster used for Data Analysis

The map-reduce tasks are scheduled and executed in parallel, thereby increasing the speed, efficiency of the entire system.

Performance of cluster (3 nodes) is tested

Low Power Embedded Processor Exploration for Multi-Physiological Stress Monitoring

A multi-physiological stress detection system containing data acquisition by sensors, feature extraction, and machine learning classifier

The entire stress detection system which consists of feature extraction and KNN classifier is implemented on various platforms including Artix-7 FPGA, NVIDIA TK1 ARM-A15 CPU and Kepler GPU and NVIDIA TX2 ARM - A57 CPU including the Maxwell and Pascal GPU

Technology: C, CUDA, MPI (multi-threaded program which runs on multiple cores),piCUDA, Matlab, Machine Learning (SVM, KNN).

Aswini Integrated Technologies, Hyderabad, India

Software Developer Sept 2014 – July 2015

Test the quality of the Product sample and track the sample status sent to the lab for testing.

Created Users for Login, assigning roles and provided authorizations to the users.

Developed interactive user interfaces using JavaScript, Twitter Bootstrap, JQuery, PHP, HTML5 and CSS.

Designed a dynamic user interface of the web application using ASP.NET

SQL Server is used as a data base and deployed the web application in Apache tomcat.

Technology: HTML, CSS, Twitter Bootstrap, JQuery, PHP, ASP.NET and SQL.

ACADEMIC PROJECTS

A Framework for Scene Recognition Using CNN as Feature Extractor and Machine Learning Kernels as Classifier

A framework was implemented in MPI for scene recognition by using Convolutional Neural Networks (CNN) as a feature extractor and using SVM and Perceptron for Classification.

Images with large dimensions were converted into features with small dimensions and the output of the CNN was provided as input to different ML Kernels. Comparison of Serial and Parallel implementation of the frame work was performed for training and classification of various images on Bluewave cluster.

Game of New Eleusis

Card game designed using Artificial intelligence and Machine Learning Techniques.

Designed an Agent who plays the game and predicts the rule decided by the dealer based on the combinations thrown on the table

RESEARCH PUBLICATIONS

Ali Jafari, Sunil Gandhi, Sri Harsha Konuru, David Hairston, Tim Oates and Tinoosh Mohsenin “An EEG Artifact Identification Embedded Hardware using ICA and Multi-Instance Learning "

Contact this candidate