Abhinav Iyappan
https://www.linkedin.com/in/abhinavvishak
*******.******@*****.*** 646-***-****
EDUCATION
COLUMBIAUNIVERSITY M.S IN ELECTRICAL ENGINEERING, DATA-DRIVEN ANALYSIS AND COMPUTATION Relevant Coursework: Deep Learning, Machine Learning, Big Data, NLP, Computer Animation Expected Dec 2016
INDIAN INSTITUTEOF TECHNOLOGYHYDERABAD B.TECH IN ELECTRICAL ENGINEERING Relevant Coursework: Detection & Estimation theory, Convex Optimization, Image & Video Processing Graduated Aug 2015
EXPERIENCE
SOUTHERNELECTRONICS INTERN
May 2014 – Jul 2014 Chennai, India
• Analysed industrial PCB designs and optimized layout space, which is expected to deliver 20% cost reductions. Worked closely with the CEOand senior analysts.
TATACONSULTANCY SERVICES ANDROID DEVELOPMENT INTERN May 2013 – Jun 2013 Chennai, India
• Developed an Android app that takes photos of signboards in foreign languages, and then translates them into languages of the user’s choice. Various techniques in image processing and image recognitionwere used.
• Managed a teamof 3 to deliver project concepts for an internal incubator. RESEARCH& PROJECTS
DEEP LEARNINGUSINGTHEANO Jan 2016 – April 2016
TrainedDeepNeural Networks to perform recognition on the real dataset of Housing Numbers SVHN ). Achieved accuracies of 90%, withminimal training time of just 10minutes.
SPEECHRECOGNITIONWITHACOUSTICAND LANGUAGEMODELS Jan 2016 – May 2016 Built parts of the complete speech recognition process using concepts as HMMs,GMMs,n-grammodels etc. on Python COMPUTERANIMATIONPHYSICS ENGINE Aug 2016 - Dec 2016 Built a physics Engine to handle and render graphical simulations. Implemented features all the way frombasic forces like gravity and spring to complex fluid dynamics and large-scale collision detection. GMMMODEL GENERATION INOPENCL Aug 2015 – Dec 2015
Generated a speedup of 80 times by optimizing the Expectation-Maximization algorithm for GaussianMixturesModels, by exploiting parallelismon a Tesla K40c GPU.
DATAANALYSISOFDOCTOR-MEDICORP RELATIONS Aug 2016 - Dec 2016 Analyzing large datasets using Bid data frameworks like hadoop and spark to findmeaningful relationships and trends between financial transactions between health care providers like doctors andmedical corporations like pharmaceuticals . PROGRAMMING
C•C++ •MATLAB• Python • Theano •OpenCL •
Pig •Hive •HBase •Mahout • Spark • Javascript
SKILLS
Clustering • Regression • Classification • Deep
Learning • Amazon EC2 • Big Data
ACHIEVEMENTS &HONORS
• Tesla Scholar honoree for outstanding academic achievement.
•University team captain for Table Tennis, IIT