Siri Raavi

siri@raavi-36SN:~$ Data Scientist with 3 years of experience mining, refining, and modelling real-world data. Also, collaborative and enthusiastic team player dedicated to efficiently solve problems. _

Experience

Data Scientist

Sweeten New York, NY OCT 2018 - Present
  • Developed a machine learning algorithm for our core operations that matches the client project with the best general contractors which improved sales by 32%.
  • Designed and developed solutions that assess the quality of the leads and predict sales for the month. Created effective workflows that reduced manual effort by 46%.
  • Built interactive mission-critical dashboards that guide day-to-day operations, planning, and strategic decision-making.Designed and developed solutions that assess the quality of the leads and predict sales for the month. Created effective workflows that reduced manual effort by 46%.
  • Set up, maintained, streamlined and scaled our data analytics infrastructure and order our data pipelines to enable machine learning at scale.
  • Collaborated with product, marketing, and sales teams translating business constraints and product management queries into experimental design and considered ways to test, track, measure, and optimize across systems.

Research Assistant

HULA Lab University of Houston Houston, TX JAN 2017 - OCT 2018
  • Developed a framework to train neural networks with memory capacity in classifying images using small number of samples as part of my Master Thesis.
  • Trained Memory Augmented neural network on MNIST data with few samples in One vs All approach for 100,000 episodes and achieved highest accuracy of 96.2% even with the presence of label noise.
  • Collaborated design and development of Generative adversarial model to retrieve chest radiographs for radiology toolkit.

Application Software Developer

NTT DATA Banglore, India JUL 2014 – JUL 2016
  • Managed the access to users and ensured optimal performance of databases and their associated objects by executing various batch jobs
  • Streamlined major implementations like performing conversions and production issuances.
  • Developed an internal tool automating few tasks reducing the manual effort and time spent on the task by 85%

Teaching Assistant

College of Technology University of Houston Houston, TX AUG 2016 – JUN 2018
  • Assisted undergraduate students in performing their experiments and implementation of their final projects.
  • Guided the students in preparation of professional. reports for the projects undertaken by them.
  • Supervised the administrative duties of the College of Technology like ABET reporting and lab management.

Skills

  • Machine Learning
  • :
  • Sci-kit Learn
  • NTLK
  • Spark-MLLib
  • Numpy
  • Pandas
  • Scipy
  • Matplotlib

  • Deep Learning
  • :
  • LSTM
  • CNN
  • RNN
  • Tensorflow
  • Theano
  • Keras
  • OpenCV
  • Lasagne

  • Programming Languages
  • :
  • Python
  • Linux
  • SciKit-learn
  • SQL
  • Flask
  • Dash
  • MATLAB

  • Tools & Utilities
  • :
  • Git
  • GCP
  • AWS
  • Azure
  • Looker
  • Excel

  • IDE/Editors
  • :
  • PyCharm
  • IntelliJ
  • Vim
  • Visual Studio
  • Eclipse
  • NetBeans

  • Software Engineering
  • Data Visualization
  • Probability and Statistics

Education

University of Houston
Master of Science (w/Thesis), Computer Systems Engineering
AUG 2016 - MAY 2018

Jawaharlal Nehru Technological University
Bachelor of Technology, Electrical and Electronics Engineering
AUG 2010 - MAY 2014

Portfolio

Radiologist Gaze

Collaborated with radiologists from M.D. Anderson Cancer Center to collect and analyze radiograph gaze pattern analysis. Gaze features were extracted using clustering and warping methods. Developed a model using random forest to capture the cognitive components of the radiological processes such as visual, attentional and decision.

ECSSGAN

Designed a generative adversarial machine learning model based on the SALGAN model and used it to detect the objects and generate their saliency maps, especially of the images that included depth information of extended complexity scenes using machine learning framework Theano and Python

Machine learning using cloud services

Evaluated the three elite cloud service providers; Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft Azure for their machine learning services by comparing the different metrics on three different UCI repositories

Gender Classification from blog text

Built a toolbox to identify the gender of the author of the text. The tool box was evaluated to predict the gender with 82% accurate results.