Lead Data Scientist

SRINIVASAN SEMBAKKAM RAJIVELU, CSM

**** ***** ******, ***. *** 514-***-****

Thornhill, L4J 1V9 ac0y0n@r.postjobfree.com

PROFILE

Having 7 years of experience in doing Statistical Data Analysis and Machine Learning.

Having experience in doing Huge Scale Data Analysis and machine learning using Spark.

Having experience working in Agile teams and a Certified Scrum Master (CSM).

Having profound knowledge in Machine Learning techniques

Having experience in developing Deep Learning Models using TensorFlow.

Having experience in Natural Language Processing and used Python NLTK.

Having experience in doing data analysis in different domains such as Insurance, Investment Banking, Advertising, Computer Vision, Social Media, Internet of things etc.

Having experience in working with SAS Enterprise Miner and developed workflows for user retention model prediction for Intact Insurance.

Having experience in doing visualizations using Tableau and PowerBI.

Having experience in developing Highly Scalable Cloud Platforms using NoSQL technologies like Hadoop, Hive and HBase.

Having experience in doing Risk modeling in Investment Portfolios across various segments.

Participated in Windows Azure Cloud Competition “Imagine Cup 2012” and my team was Semi Finalists in the competition.

Maintain focus in demanding work environments, under deadline and pressure conditions.

Good communication and interpersonal skills. Excellent adaptability to new technologies and new domains.

EDUCATION

Master of Software Engineering

Concordia University, Montreal, Quebec March 2013

(Specialization: Machine Learning and

Natural Language Processing)

Bachelor of Computer Science

Anna University, Chennai, India 2003 – 2007

TECHNICAL SKILLS

Operating System: Windows and Linux.

Programming and Scripting Languages: Python, Scala and R

Machine Learning Libraries: Spark MLLib, TensorFlow, Spark Streaming, Spark SQL

Big Data Technologies: Kafka, Hadoop, SQL, HBase and Hive

Data Visualization: Tableau and PowerBI

CERTIFICATIONS

CERTIFIED SCRUM MASTER (CSM)

Microsoft Professional Program Certification in Data Science (In Progress)

Research Works:

Natural Language Processing Lab – Concordia University (June 2011 – September 2011)

Social Networking – Sentiment Analysis and Spam Detection

Developed a system, which can take Tweets from the twitter for a specific keyword and can give positive tweets and negative tweets for the keyword.

Analyzed the background information about the user and detect the user is a spam or not.

Social Networking – Topic Modeling

Developed a system, which can take Facebook feeds and do a topic modeling using Latent Dirichlet Allocation (LDA) Algorithm.

Semantic Wiki Model for Requirements Engineering

Developed Semantic Web Forms on Wikipedia with that user can enter Software Requirements and requirements are analyzed and related using the Semantic Wiki Model. An IEEE Poster has been made out of this project for WikiSym 2012.

And this project has been integrated as part of the research thesis titled ”A General Architecture to Enhance Wiki Systems with Natural Language Processing Techniques” to evaluate the architecture.

Cloud Computing Lab – Big Data Researcher, Concordia University (October 2011 – June 2014)

Meetspot – Social Location Recommendation Engine

Developed an application called Meet Spot, which recommends the locations based on User Check-in Details.

Meetspot recommendation system considers both frequency and social aspects as the criteria for the recommendation.

Developed for Concordia Poster Competition. Won the First prize for this application.

Publications

Sembakkam Rajivelu S, Meet Spot - A Meeting Spot Recommendation System based on Geo Location Data in Social Network, ECGSA Concordia Poster 2012.

Sembakkam Rajivelu S, User Interest Evolution in Social Data, ECGSA Concordia Poster 2012.

Sateli B., Sembakkam Rajivelu S., Anguis E., Witte R., "ReqWiki: A Semantic System for Collaborative Software Requirements Engineering", WikiSym '12, Linz, Austria.

Sateli B., Anguis E., Sembakkam Rajivelu S., Witte R., "Can Text Mining Assistants Help to Improve Requirements Specifications?" MUD '12, Kingston, Ontario, Canada.

CAREER RELATED EXPERIENCE

Deep Insights Inc., Toronto Canada

Lead Data Scientist Consultant (Jan 2016 – Till now)

User Retention Model

Worked for Intact Insurance “Actuarial Team Belairdirect Quebec” on User retention Model with SAS EG and SAS EM. Improved and beat their baseline model performance by 11% with machine learning techniques.

Tested different machine learning algorithms and done evaluation of the model using Lift Curve and Confusion Matrix and presented the results to the business people.

Stock Prediction Model

Worked for Intact Insurance “Investment Management Team”. Improved and beat their stock prediction model performance by 12% with machine learning techniques.

Some of the algorithms tested in this model include Logistic Regression, Support Vector Machines, Gradient Boost, Decision Tree, Random Forest and Artificial Neural Network.

Tested different techniques for feature extraction and identified top 23 features out of 800.

Tested the model by doing experiments like rolling window and expanding window with different time slices in the data.

Investment Portfolio Risk Model

Developed risk models on Capital Markets Investment Portfolio using techniques like Monte Carlo Simulation etc.

Assessed the risk and modeled different risk models on portfolio’s across different segments and presented the results to the stakeholders.

Anomaly Detection Model

Worked on Anomaly Detection Model using DBSCAN algorithm.

Successfully clustered payment transaction data using KMeans, DBSCAN and hierarchical clustering methods.

Evaluated the Model and fine tuned to improve performance.

Developed different dashboards using Tableau and PowerBI.

User Interest Prediction – Deep Learning

Applied Deep Learning word2Vec model on Facebook feeds.

Produced word vectors i.e User Interested Topics with deep learning via word2vec’s “skip-gram and CBOW models”, using hierarchical softmax sampling techniques in Facebook feeds.

Connected the interested topics with users and developed a user interest prediction model.

Fashion Item Detection – Deep Learning

Developed a deep learning model for fashion item detection in images.

Developed a Faster recurrent convolution neural network using Tensorflow for doing fashion item detection in images.

Cossette, Montreal Canada

Big Data Analyst (July 2014 – Dec 2015)

Data Warehousing Solution

Developed a data warehousing solution using Hortonworks 2.3 platform in Amazon EC2.

Deployed Big Data Technologies like Kafka, Hadoop, Hive and HBase.

Exported data from MSSQL database to Hadoop using Sqoop.

Developed ETL pipelines to get data from different systems like Facebook Analytics, Google Analytics, Adobe Campaign Logs and Mobile App Logs.

Developed ETL solutions with spark streaming and Spark SQL for real time data analytics.

Digital Advertisement Analytics

Developed analytics pipelines on digital advertisement data using python Scikit learn and Spark MLLib.

Analytics on digital advertisement includes churn prediction, finding active/passive users, channel prediction, advertisement recommendation etc.

Recipe Recommendation System

Developed a recommendation system, which recommends recipes based on user profile interests for alzheimer patients.

Recommendation approach used here is case based recommendation.

Luci – A virtual assistant for Alzheimer patients

Luci is a personal virtual assistant for Alzheimer patients, which recommends food items, games based on patient profile.

I have developed the back end for this project using MSSQL, HBase and Hadoop.

Developed a question-answering framework using Python.

Based on the answers, appropriate data will be queried, rules will be applied and contents will be present in the mobile app.

Mnubo, Montreal Canada

Big Data Architect (Nov 2013 – June 2014)

IOT Data Analytics Platform

Worked on predictive analytics on Internet of things (IOT) data in Hadoop using Apache Spark MLLib and R.

Developed REST API using Nodejs Express for real time alerts and notifications.

Merchlar, Montreal Canada

Data Scientist Intern (April 2013 – October 2013)

Developed a backend system using Hadoop and HBase to store the logs generated by Merchlar Computer Vision SDK.

Developed real time analytics pipelines using Spark Streaming.

Done predictive analytics using Apache Spark MLlib and R.

Concordia University, Montreal Canada

Teaching Assistant (Jan 2011 – March 2013)

Worked as a Tutor for courses such as Advanced Programming in C++, Data Structures and Algorithms in C++, Tools and Techniques in Software Engineering etc.

Global Edge Software Pvt Ltd, Bangalore, India

Software Engineer (June 2009 – March 2010)

Worked as C++ Consultant for Infineon Technologies, Germany for 3 Months.

TATA ELXSI Pvt Ltd, Bangalore, India

Software Engineer (June 2007 – June 2009)

Worked as a C++ Developer for VOIP Applications like IpPhone, Video Phone.

Responsibilities include design, development of User Interface for Desktop applications.

Contact this candidate