Sign in

Machine Learning Engineer

Amherst, Massachusetts, United States
April 05, 2018

Contact this candidate



**F, Brandywine Drive, Amherst, MA 01002 Phone: 408-***-****

Work history

NLP Research Intern May 2017 - Sep 2017

Information Extraction and Synthesis Laboratory (IESL) Amherst, MA

Worked under Prof. Andrew McAllum on Extractive Single-Document Summarization via tree constrained inference and recursive cardinality potentials on NYT corpus in TensorFlow

Coded a transition-based dependency parser in PyTorch for scientific literature

Created BLESS datasets in multiple languages for unsupervised hypernym detection

Associate Applications Developer Sep 2015 - Jun 2016

Oracle Financial Services Software Limited Mumbai, India

Programmed REST API based web services for banks following agile methodology, designed the UI, handled the life cycle of User-Management module and orchestrated a penetration testing of the application


Master of Science: Computer Science Sep 2016 - May 2018

University of Massachusetts Amherst Amherst, MA

Coursework – Deep Learning, Advanced NLP, Machine Learning, Probabilistic Graphical models; GPA- 3.83

Minor in Data Science – Algorithms for Data Science, Systems and Database design

Teaching Assistant – Advanced Machine Learning and Secure Distributed Systems (Blockchain)

Bachelor of Engineering (Honours): Electronics and Communication Aug 2011 - July 2015

Birla Institute of Technology and Science (BITS Pilani) Hyderabad, India

Internship – Front end Developer at MaaS360, an IBM Company;


Coding – Python, Java, PostgreSQL, MATLAB, C, JavaScript

Tools – TensorFlow, PyTorch, NumPy, Scikit-learn, Pandas, Keras, NLTK, AWS, Linux, Git, PyCharm


1)Improving Open Domain Dialogue-Systems (Chatbots)

Built an seq2seq neural conversational model in PyTorch using attention with intention and a diversity promoting objective function to prevent irrelevant boring outputs

2)Multilingual embeddings for cross-language NLP

Created language agnostic word embeddings via Artificial Code-Switching to run any NLP task for languages with very less labeled data

3)DSR Reinforcement Learning to navigate a Labyrinth

Applied successor representations within an end-to-end deep reinforcement learning framework, comparing its efficacy to DQN on a grid-world domain (Mazebase) on AWS EC2

4)Sarcasm Detection for Target-dependent Sentiment classification

Trained a model on twitter data to detect sarcasm using n-grams, sentiment scores and topic modeling (LDA) and used it to improve sentiment classification of Amazon reviews


First place in Google Hackathon - Used Firebase and Google maps API to develop an app, InTheBin! that renders crowd-sourced locations of nearby trash cans

Contact this candidate