Amol Shripad Takkalki

* (Lotus), Five Flowers

Near Chatrapati Bank, Jagtap dairy,

Pune, Maharashtra, India 400027

+91 814-***-****

ab9j9o@r.postjobfree.com

Objective

To obtain a job that utilizes my skills in my area of interest.

Professional Interests

Data analysis, Computational statistics, Machine learning, Parallel Computing and Hadoop.

Professional Experience

Michigan Technological University, USA (1.5 years) Houghton, MI, USA

Research Assistant at Department of Computer Science August 2011-December 2012

Education

University of Pune Pune, India

M.Tech Modeling and Simulation 2010

University of Pune Pune, India

B.E Chemical Engineering 2008

Technical Skills

Professional Business Experience

Operating systems: Linux and Windows

Languages: C, R, MATLAB/OCTAVE, Perl, C++ .

Software: Aspen plus, Gambit/Fluent, Weka/Orange.

Libraries: Xlib, GSL (GNU Scientific library), libsvm, MPI, CUDA/GPU.

I have used C and ‘MATLAB’ extensively for numerical methods.

Most of the programs which involved numerical methods were written using the GSL library.

I have used ‘R’ extensively for computational statistics and evolutionary algorithms.

I have also written five technical articles for the magazine ‘Linux For You’ on GSL and Xlib.

Single author research paper published at “International conference of computing”, New Delhi.

Currently, I am learning Microsoft Project, Hadoop and SQL.

Language Skills

Environmental Engineering Chemical Processes Laboratories

Written and Spoken Fluency in English, Hindi and Marathi

st

Last updated: July 1, 2013 Amol S. Takkalki, Page 1 of 2

Projects

Michigan Technological University

Research assistant (1.5 years): Research project

Advisor: Dr. Laura Brown

Language/Software: MATLAB

Objective: To create a system/algorithm to fill in the missing values of data.

Data features: time series variable data, 11k patient instances.

University of Pune

Research assistant (7 months): Research project

Language/Software: MATLAB

Objective: Generate a model for emission readings of fuel-biofuel mixture from an engine.

Data features: Multi-dimensional output data, sparse data set,

The first objective was to create a model using neural networks to predict the emission for an

unknown blend.

The second objective was to find the best blend theoretically and then validate the results

experimentally (not done by me).

Michigan Technological University

Course project: Advanced artificial intelligence

Advisor: Dr. Laura Brown

Language/Software: R, Orange

Objective: Prediction of baseball players’ salaries based on some observed instances.

Data features: Regression type data.

A number of models were created and the best model was selected based on value of least mean

squared error. Each model consisted of a different set of features that were selected using genetic

algorithm.

Michigan Technological University

Course project: Computational statistics

Advisor: Dr. Andreas

Language/Software: R, Orange

Objective: Comparison of results generated by Genetic Algorithms and Simulated Annealing algorithm

Data features: Regression type data, 9k instances.

Feature set was first selected using ranking methods and both algorithms and then the final model

was compared along with the evolution of the algorithm.

University of Pune

Course project: Machine learning

Language/Software: C, libsvm (Support vector machine’s C library)

Objective: To create a model for fault classification

Data features: Class type data, control system data, more than 100 attributes

Several models were created using well known supervised learning algorithms and one of the

models with highest prediction accuracy on test data was reported to be the best one.

University of Pune

Course project: C programming

Language/Software: C

Objective: To generate a market analysis model

Data features: Regression type, supply-demand data, two-dimensional data

For this problem, we had taken real time data of Australian electricity market. The relation between

supply and demand was appropriately modeled using regression methods and the results were

predicted accordingly.

st

Last updated: July 1, 2013 Amol S. Takkalki, Page 2 of 2

Contact this candidate