About Me


        

  Professional Summary

Analytics Specialist

Opera Solutions

Noida, Uttar Pradesh,     Jul 2015 - Presently WorkingCurrent Experience


    

Visiting Research Assistant

Nanyang Technological University

Singapore, Singapore,     May 2014 - Jul 2014Previous Experience

Deep Neural Networks based Voice Activity Detection

Research Intern

Philips

Bengaluru / Bangalore, Karnataka,     May 2013 - Jul 2013Previous Experience

Predicting the Needle Guidance Path for 3D Ultrasound Medical Image Analysis

  University Education and Certifications

  My Projects

Accelerated Risk Management System for Audit Engagements

Jul 2015 - Jul 2016 (1 years 0 months)

Building a risk measurement system to assess engagements for professional services company Objective of the work is to develop a model to classify specific sentences in official engagements from over 40,000 documents into safe or risky. At an overall level, predict the anomaly index of the document and construct a feedback decision system for reviewing an engagement

Student Response Analysis using Textual Entailment

Sep 2013 - Nov 2013 (0 years 2 months)

To recognize the extent of correctness of a student answer given a question and a few reference
answers, used a combination of overlap measures and semantic similarity metrics utilising parse
trees, WordNet hierarchies and Explicit Semantic Analysis which represents text as a vector of
Semantic concepts learnt using LSA. The obtained accuracy was as good as the team positioned third in the Semeval 2013 task.

Recognizing Hand Written Characters and Numerals

Sep 2013 - Nov 2013 (0 years 2 months)

To classify handwritten digits using the famous MNIST data consisting of handwritten images of English alphabets and numerals. Deployed standard machine learning classifiers such as SVM, Neural Nets, k-NN and hybrid classifiers like Ada Boost on extracted dominant features to decrease error, given limitation on the run time of algorithm.

Voice Activity Detection using Deep Belief Networks by Source and System Level Information Fusion

Jul 2014 - Jun 2015 (0 years 11 months)

Built efficient Voice Activity Detection (VAD) models using Deep Belief Networks for classifying speech signal at very low SNR (<5dB) levels. Established the importance of fusing laryngeal originated source features with features from vocal tract in training a VAD model. Various experiments were performed with several feature sets. The results were evaluated in non-stationary noise conditions and compared with state-of-the-art VAD algorithms.

  Awards& Honor

  Research, Authorship and Publications

 

 

  Blog Posts


Opportunities

Looking for job / talent?

Access Marketplace

Chance to earn up to 1000$ / referral

Refer & Earn

Similar Profiles

Rahul Kumar

Software Developer at self employed

Noida, Uttar Pradesh, India
Anita Pal

b.tech final year at student

Noida, Uttar Pradesh, India
aditya kumar

Lead Engineer at HCL Technologies

Noida, Uttar Pradesh, India
surjeet negi

Ab Initio Developer at John Lewis Partnership

Noida, Uttar Pradesh, India
Shrey Srivastava

BTech in Computer Science and Engineering at AKTU,Lucknow

Noida, Uttar Pradesh, India
Huda Hasan Rizvi

Digital marketer at KAZO Fashion Ltd

Noida, Uttar Pradesh, India
Sahil Khurana

Software engineer at HCL Technologies

Noida, Uttar Pradesh, India
Ashish Gupta

Business Analyst at G4S Corporate Services Pvt.Ltd.

Noida, Uttar Pradesh, India
Aman Dubey

Sr. Executive at Idea Cellular Limited

Noida, Uttar Pradesh, India
Pratik Vikash Singh

Noida, Uttar Pradesh, India