Deeksha Doddahonnaiah

MSCS Gradute Student

Northeastern University

“It’s not magic, it’s machine learning”

Hi! I’m Deeksha. My favorite programming language is R! I find statistics, math and algorithms fascinating, and inevitably Machine Learning is the field which interests me the most.

In my free time, like to read, usually about data science, neuroscience or philosophy. Among other thing I enjoy biking, weight training and drawing.

Interests

Machine learning
Computational Biology
Data Mining
Statistics

Education

Master of Science in Computer Science, 2019

Northeastern University
BE in Computer Science, 2016

Sapthagiri College of Engineering

Skills

R

tidyverse

Python

sklearn

Visualization

D3, javascript

Experience

Graduate teaching assistant

Khoury College of Computer Science

Jan 2019 – Present Boston

Graduate teaching assistant for Data mining techniques/Unsupervised ML (CS6220)

Computational Biologist Co-op

Alnylam Pharmacueticals

Jul 2018 – Dec 2018 Cambridge

Performed data extraction and manipulation on LINCS L1000 mRNA assay to investigate the effect of perturbations on gene-expression.
Applied statistical methods like regression analysis and correlation tests, and visualization techniques using R and tidyverse to find interesting trends with respect to gene dysregulation.
Experimented various regression models like Neural Regression, Random Forest Regression and Linear Regression to find patterns in large public datasets.
Developed a web application using Flask to add, list and monitor plasmid AAV orders.

Reaserch Assistant

Mass Open Cloud

Jan 2018 – Apr 2018 Boston

Deploying Multnode Openstack on a kaizen environment, using Packstack and Devstack.
Preparing a testbed with tracing tools such as OSProfler and Rally to analyze performance of the nodes.

Solution Specialist

Comptel Corporations(Acquired by Nokia)

Jun 2017 – Jun 2018 Bengaluru

Developed Network Element Interfaces and Business Service Logics for vendors like Telsra, TCL and Idea to name a few.
Worked on the Activation and Provisioning lower down in the telecom stack. InstantLink/FlowOne being the product used to develop logic and facilitate the flow of requests from OSS/BSS to Network element Interface.

Accomplishments

Presenter at NUViz

Northeastern University Visualization Consortium(NUViz) Apr 2019

Presented MURA Classification tool

Winner of logo design contest

Alnylam Pharmacueticals Aug 2018

Designed the logo for Alylam Research Department

Neural Networks and Deep Learning

Coursera Apr 2018

Projects

MURA Classification Tool

Visualizing the classification results of MURA using D3 and javascript. Implemeted multiple CNNs to classify body parts(AlexNet) and abnormality for a specific body part(DenseNets)

PDF Video Code

Stock Analysis using Topic Modeling(LDA)

Study the impact of news headlines across various topics on NASDAQ prices. Employed and compared performace results of topic modeling(LDA), regression analysis, and neural networks.

PDF Code

Classification methods for Predictive Credit Scoring

Study and analyze classification methods for credit-scoring. Classifier used : SVM, Random forests, and Neural Networks.