Sunanda Bansal

Artificial Intelligence Solution Architect
Data Scientist
Machine Learning Engineer
Natural Language Processing Specialist

GitHub: github.com/sunandabansal
LinkedIn: linkedin.com/in/sunandabansal

Contact me: Google form

Summary

MS (Thesis) in Computer Science, with 4 years of experience conducting research in Natural Language Processing (NLP) with Machine Learning (ML).
5+ years of experience of engineering and successfully delivering 15+ AI solutions.
5+ years of experience with Python, 5+ years of experience with Linux, total 7+ years of experience in software/web development.
Track record of significantly improving performance in AI solutions.

Work History

DATA SCIENCE CONSULTANT, 3.5 years
Deloitte, Dataperformers (Acquired by Deloitte) (Jun. 2019 – Nov. 2023)

AI Solution Development - Led the design and development of 15+ diverse AI-driven solutions across various domains, including tagging and classification, search and ranking systems, chatbot solutions in response to COVID-19 challenges and more.
Performance Optimization and Accuracy Enhancement - Significantly improved performance in various AI projects, elevating accuracy and recall rates from initial scores to impressive levels often within tight deadlines. (Example – F1 Score from 32% to 92% in 2 days)
Product Development and Project Revitalization – Conceptualized products and resurrected 5+ internal tools demonstrating expertise in project management, cross-team management, and leadership.
Strategic Leadership – Planned enhancements and spearheaded discussions on future steps, fostering collaborative dialogue and delivering value-driven and client-centric solutions.

Skills/Technologies: ML, NLP, Cloud (GCP), Web Dev, UX/UI, Scikit-learn, Gensim, SpaCy, Keras, Docker, Flask, React, Django.

RESEARCH ASSISTANT, 2 years
Computational Linguistics Lab at Concordia (CLaC) (Sep. 2017 – Apr. 2019)

Text Document Representation Research - Led pioneering research in text document representation, resulting in the development of Word Cluster based Document Embedding (WcDe), demonstrating its superior performance across diverse topics in over 14,000 experiments.
Sentiment and Emotion Analysis – Analyzed and estimated affect in texts including emotion classification, intensity estimation, and valence assessment in tweets as well as dialogue-based emotion analysis.
Feature Engineering with Computational Linguistics – Applied computational linguistics to engineer statistical-linguistic features like based on analysis of dependency parse trees and constituents, enriching Machine Learning models for sophisticated text analysis.
Development and Integration – Streamlined development and integrated various linguistic analysis tools in the lab’s text analysis software (GATE) for direct use in processing pipelines.

Skills/Technologies: Emotion Analysis, Document Classification, Topic Modelling, LDA, LSA, Doc2Vec, BERTopic, Word Embeddings, Document Embeddings, Word2Vec, GloVe.

TEACHING ASSISTANT, 2 years
Concordia University (Sep. 2017 – Apr. 2019)

Tutored Discrete Mathematics, Artificial Intelligence, Intelligent Systems, and Information Retrieval.
Engaging Tutorial Sessions - Provided valuable individual and group assistance, fostering a collaborative learning environment.
Lectures and Lab Demonstrations - Tutored sessions and demonstrated labs, evaluated assignments and projects.
Course Enhancement Contributions - Prepared specialized exercises on Natural Language Processing (NLP) for the Artificial Intelligence class, enhancing the curriculum.
Effective Class Management - Managed and taught up to 3 classes each term, demonstrating organizational skills and the ability to handle multiple responsibilities.
Collaborative Decision-Making - Assisted professors with decision-making processes, actively contributing to the improvement of course structures and strategies.

FOUNDING ENGINEER / DIGITAL DESIGN ARCHITECT, 2 years
Poplify / Intuzion Technologies (Jan. 2013 – Nov. 2014)

Graphic and Web Development - Proficient in Graphics Development, Print and Web Design, and Branding, demonstrating a strong aesthetic sense and technical proficiency.
Frontend Web Development Specialist - Specialized in front-end web development, focusing on E-commerce Web Apps and Content Management Systems, ensuring optimal user experiences.
International Recognition - Competed against international brand designers and won an international competition for logo design of a personal trainer’s brand. Also recognized in the top 3 of several logo design competitions for various brands and organizations.
Versatile Designer - Demonstrated expertise in a variety of industries, from luxury restaurants to e-commerce platforms, combining creativity and technical skills.
Team Collaboration - Worked seamlessly with international teams, providing design solutions while ensuring consistency and effective communication.
Problem Solver - Successfully addressed challenges in tight deadlines, new frameworks, and diverse design requirements, showcasing adaptability and resilience.

Skills/Technologies: HTML, CSS, SCSS, jQuery, JavaScript, CMS, Ruby, Ruby on Rails, PHP, PHP Codeigniter.

Education

Master of Computer Science (Thesis), Concordia University, Montreal (Sep. 2017 – Aug. 2021)
Thesis – Vector Representation of Documents using Word Clusters

Bachelor of Engineering in Computer Science (Hons.), Panjab University (Jul. 2009 – Jun. 2013)

Publication

Vector Representation of Documents using Word Clusters
Graduate Thesis · Concordia University · Aug 2021

For processing the textual data using statistical methods like Machine Learning (ML), the data often needs to be represented in the form of a vector. With the dawn of the internet, the amount of textual data has exploded, and, partly owing to its size, most of this data is unlabeled. Therefore, often for sorting and analyzing text documents, the documents have to be represented in an unsupervised way, i.e. with no prior knowledge of expected output or labels. Most of the existing unsupervised methodologies do not factor in the similarity between words, and if they do, it can be further improved upon. This thesis discusses Word Cluster based Document Embedding (WcDe) where the documents are represented in terms of clusters of similar words and, compares its performance in representing documents at two levels of topical similarity - general and specific. This thesis shows that WcDe outperforms existing unsupervised representation methodologies at both levels of topical similarity. Furthermore, this thesis analyzes variations of WcDe with respect to its components and discusses the combination of components that consistently performs well across both topical levels. Finally, this thesis analyses the document vector generated by WcDe on two fronts, i.e. whether it captures the similarity of documents within a class, and whether it captures the dissimilarity of documents belonging to different classes. The analysis shows that Word Cluster based Document Embedding is able to encode both aspects of document representation very well and on both of the topical levels.