Akshay Kalia

Akshay Kalia

New York, New York, United States
689 followers 500+ connections

About

Experienced technologist, architect, and leader. Full-stack software developer with more…

Activity

Join now to see all activity

Experience

  • Pillar Technologies, Inc. Graphic

    Pillar Technologies, Inc.

    Greater New York City Area

  • -

    Greater New York City Area

  • -

    Sunnyvale, California

  • -

    Gurgaon, India

  • -

    Gurgaon, India

Education

Courses

  • Computer Vision

    CSCI-GA 2271

  • Data Mining for Business Analytics

    INFO-GB.3336

  • Fundamental Algorithms

    CSCI-GA.1170-​001

  • Graphics Processing Units (GPUs): Architecture and Programming

    CSCI-GA 3033

  • Natural Language Processing

    CSCI-GA.2590-​001

  • Operating Systems

    CSCI-GA.2250-​001

  • Programming Languages

    CSCI-GA.2110-​001

  • Realtime & Big Data Analytics

    CSCI-GA.3033-​006

  • Web Search Engines

    CSCI-GA 2580

Projects

  • Predictive Model for Cancer detection using non-invasive methods

    Cornell Health Hackathon 2016 Finalist

    - Built a model using various machine learning algorithms to predict the type of cancer using a bio-marker in blood plasma or serum using the technology of "liquid biopsy"

    - Worked on a dataset comprising 673 samples with a feature space of 1046 to detect Pancreatic and Ovarian carcinoma based on a bio-marker using Machine Learning algorithms like Linear SVM, Gaussian SVM, Logistic Regression, Naive Bayes, etc.

    Stack : Python…

    Cornell Health Hackathon 2016 Finalist

    - Built a model using various machine learning algorithms to predict the type of cancer using a bio-marker in blood plasma or serum using the technology of "liquid biopsy"

    - Worked on a dataset comprising 673 samples with a feature space of 1046 to detect Pancreatic and Ovarian carcinoma based on a bio-marker using Machine Learning algorithms like Linear SVM, Gaussian SVM, Logistic Regression, Naive Bayes, etc.

    Stack : Python, scikit-learn

  • Image (JPEG/PNG) Compression Algorithm

    Designed an Image (JPEG/PNG) Compression Algorithm, without using any 3rd party libraries, which can compress photos without any visible loss in quality or dimensions. Currently working on video compression using the same algorithm.

    Stack : ASP.NET 4, C#, HTML5, CSS, jQuery

    See project
  • Chrome Spell Check Plugin

    Developed an auto-correct Chrome Extension, which has been given good reviews by leading internet blogs.

    Stack : jQuery

  • Funny Image Website

    Developed funny images sharing website, which has currently over 3.5 million page views per month and active user base. Images are automatically added to the website with help of crawlers and user submissions.

    Stack : ASP.NET, C#, HTML5, CSS, jQuery

  • Question Answering (QA) System based on Natural Language Processing

    -

    This project deals with building a Question Answering system in Python using NLP techniques like POS, BIO, NER tagging, lexical element extraction using our own algorithms without the use of any external libraries like NLTK and an IR module - SOLR.

    The performance of a few modules of the system are as follows:
    BIO Tagger: F-1 score of 95.08
    NER Tagger: F-1 score of 81.00
    Question Classification: accuracy of 80%

    - Built POS tagger and trained it on a tagged Wikipedia dump…

    This project deals with building a Question Answering system in Python using NLP techniques like POS, BIO, NER tagging, lexical element extraction using our own algorithms without the use of any external libraries like NLTK and an IR module - SOLR.

    The performance of a few modules of the system are as follows:
    BIO Tagger: F-1 score of 95.08
    NER Tagger: F-1 score of 81.00
    Question Classification: accuracy of 80%

    - Built POS tagger and trained it on a tagged Wikipedia dump and generated the model file.
    - Built BIO tagger using bigram as well as trigram feature conjunction and trained it against the Wikipedia dump. The tags are assigned using Maxent model.
    - Built NER tagger using bigram feature conjunction, external dictionaries and word semantics and trained it against the Wikipedia dump.
    - Built Feature File for Question/Answer Type classification and trained it against the TREC dataset.
    - Performed stemming and Keyword Extraction
    - Did Information Retrieval using SOLR
    - Summarized the response from content in Wikipedia which is to be given back to the user

    Stack: Python, SOLR

    Other creators
  • Cluster Analysis on Restaurant Data to Produce a Rating/Review on Google Maps API

    -

    The analytic grabbed data from Yelp, Google and Zomato and cleaned the data using hive. Analyzed data using mapper and reducer functions, broke it down to different neighborhoods and thus generated a rating for that neighborhood. The result was then plotted on Google Maps using the API.

    Stack : Java, Python, Hadoop, Hive, Google API

    Other creators

Test Scores

  • TOEFL

    Score: 112/120

  • Graduate Record Examination (GRE)

    Score: 323/340

Languages

  • English

    Full professional proficiency

  • Hindi

    Native or bilingual proficiency

Recommendations received

More activity by Akshay

View Akshay’s full profile

  • See who you know in common
  • Get introduced
  • Contact Akshay directly
Join to view full profile

Other similar profiles

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Others named Akshay Kalia

Add new skills with these courses