Mainly focused on Machine Learning application for NLP
Knowledge Base extracted from Wikipedia Articles [Graph database used]
Create a Knowledge base using Wikipedia in the form of RDF triplet and then use BigData Query Language to fire queries over the Knowledge Base to extract relevant information. It uses Information Extraction and NER to find the triplets in the form (Subject,relation,Object). It used Graph Database for Storage and Query Pipeline.
BigData Query Language in Mesos/Spark [Ad-hoc query system]
Query Wikipedia to extract information using a knowledge retrieval query language (similar to Pig Latin) on Mesos and Spark (setup on EC2). Data was unstructured and queries were fired to extract information not present in labeled form.
Semi-Supervised Multiclass Labeling using Laplacian Eigenmaps and Dimensionality Reduction
Implemented an idea from a research paper for propagating labels over the datasets for Multi-Class labeling using modified Laplacian Eigenmaps and Dimensionality Reduction.
Real Time Collaborative Document Editing (server back-end in PHP)
To collaboratively edit a single document from different client connection in Real Time. Simultaneous editing was done by using a lock mechanism to prevent users to edit the locked block. Each user was allowed to edit a single block at a time.
Expert System simulation for canal management
Used CLIPS language to simulate ships movements through cross-Florida canal and optimized it using heuristics. I have to use rules to optimize the total throughput of all ships individually
Interpreter and Compiler for a functional language
Designed a functional language Interpreter in C++ and also a Compiler to compile the program to native MIPS assembly.
Python, Java, groovy, C/C++, Clojure,Scheme, Go
Logic Programming, Probabilistic Logic, NLU, Knowledge Graph
Linux, Windows NT
Torch7, TensorFlow, python/numpy/nltk/scikit-learn, Julia, R, Matlab
Cloud Computing Engine:
Hadoop, Google App Engine, Heroku, AWS, FutureGrid, Mesos/Spark, Disco/python
git, svn, elasticsearch, lucene,
Resumes, and other information uploaded or provided by the user, are considered User Content governed by our Terms & Conditions. As such, it is not owned by us, and it is the user who retains ownership over such content.
Many factors go into creating a strong resume. Here are a few tweaks that could improve the score of this resume:
Job Titles Held: