Shyam Varsani

As a results-driven ML Engineer, I am proficient in handling structured and unstructured data, I’ve honed my skills in creating advanced machine learning models during my Computer Science degree at the University of Alberta. My passion lies in exploring the intricacies of NLP, where I’ve delved into sentiment analysis, named entity recognition, and language model design. Experienced in troubleshooting and risk assessment, I thrive in dynamic environments. With excellent communication and customer service skills, I collaborate seamlessly across teams, adapting to changing business demands.

Natural Language Processing - Relation Extraction & classification, HMMs & PoS Tagging,  Language Models, FSA, CFG- AI python, nltk, pandas, csv, argparse, Regex

• Made multiple grammar checkers, language models, data taggers and FSAs using different data sets(Penn Treebank, NLTK) and lexicons(General Language Lexicons). Imported datasets using SFTP from CSVs and Pre-processed it to remove noise and maintain data consistency.

Multilayer Perceptron for Hand Written Digit Recognition  & Regression for Hand Written character Recognition- AI python, TensorFlow

• Implements Linear Regression and two variations of Logistic Regression(with MSE loss and Cross Entropy Loss) for recognizing handwritten digits. Uses a multi layer Neural Network for recognizing handwritten digits. Using the MNIST data set to learn the classifiers.

Question Answer Forum (SQL & MongoDB) - Database Management Python, SQLite, MongoDB

• This program lets users ask and answer questions, search for posts and vote on posts. Privileged users can also perform other tasks such as marking a post, editing a post, and more.

Robotics - Differential drive, Omni drive robots & planar robot arm - AI python, UVS, Forward & Inverse kinematics 

• Made multiple robots using the Lego EV3 Mindstorms kit along with omni-wheels using solid works and 3D printers. Robots can autonomously localize and navigate.

Habiteer - Software Engineering Java, Android Studio, Google Firebase, Cloud Computing, SQL, CSV, SFTP

• Designed and developed an android habit tracking app with a team of 6 colleagues using Java, Android Studio, and Google firebase. As part of the same course, integrated data from multiple versions of the app created by different groups of students(using SQL server), and using an Enterprise Service Bus(RabbitMQ)