Mayank Anand Email & Phone Number

Toronto, Ontario, Canada

Halifax, Nova Scotia, Canada

• Collaborated with a startup on financial data processing, shredded transaction JSON and created embeddings using Azure OpenAI Embeddings further stored in Azure CosmosDB to perform Vector Search; integrated into chatbot using RAG framework.• Co-authored, master’s thesis research findings; supplemented constituency trees for MRC task. Showcased better modeling of long-distance dependencies with heightened attention to grammatical nuances; submitted at the 37th Canadian Conference on AI.

Halifax, Nova Scotia, Canada

Marking student’s resume and cover letter enrolled under Co-op Orientation Course.

Halifax, Nova Scotia, Canada

Advisor: Dr. Vlado KeseljNLP CSCI-6509• Assisted Dr. Keselj, conducted laboratory sessions, addressed student inquiries, and assisted in marking assignments.• Co-created laboratory document for Lab 8: Python Tutorial with PyTorch

Halifax, Nova Scotia, Canada

IT Helpdesk – Student navigator • Provided comprehensive technical support to students, staff, and faculty through phone, email, and in-person inquiries, ensuring timely and accurate issue resolution.• Delivered detailed instructions for application use, installation, and troubleshooting, improving user experience and system functionality.• Skillfully resolved IT-related challenges, including password and multi-factor authentication resets, and provided software installation assistance.• Assisted library patrons with navigating and utilizing resources efficiently and managing study space reservations, enhancing overall library operations.

Halifax, Nova Scotia, Canada

Advisor: Dr. Kim Davies, Amit Robert Baroi• Collaborated with Davies Biological Sciences Lab at the University of New Brunswick to automate the manual sorting of 45k shadowgraph images captured underwater in the Bay of Fundy for research purposes. Utilized computer vision and deep learning to develop an ensemble model that achieved a 94.42% accuracy rate, striking an optimal False Positive-False Negative trade-off.• Deployed trained deep ensemble of CNNs model on AWS, accelerating shadowgraph image sorting by up to 8x times and outperforming human precision in animal identification and segmentation, effectively transforming a labour-intensive task into an automated, efficient workflow.• Employed autoencoder-based unsupervised techniques for data exploration in shadowgraph image analysis. Generated hidden representations to evaluate clustering algorithms like K-means, DBSCAN, and EM, and leveraged t-SNE for cluster visualization.• Designed an animal segmentation model in PyTorch, Python and OpenCV using an NLP-inspired encoder-decoder architecture with stacked convolutions and deconvolutions. Achieved superior boundary delineation in generated masks compared to true masks, enhancing segmentation accuracy.• Optimized TensorBoard training graphs and rigorously evaluated models using metrics like Confusion Matrices, Accuracy Scores, and F1 measures. Leveraged Tableau, Matplotlib, Plotly, and Seaborn for advanced visualizations, including t-SNE plots and ROC curves, while employing Pandas and NumPy for data manipulation, elevating analytical insights.• Spearheaded project timelines and deliverables using agile methodology, achieving weekly sprint releases with clear updates. Culminated the project by authoring a blog post hosted on the DeepSense website, encapsulating key findings and methodologies.

Halifax, Nova Scotia, Canada

IT Helpdesk – Student navigator • Provided comprehensive technical support to students, staff, and faculty through phone, email, and in-person inquiries, ensuring timely and accurate issue resolution.• Delivered detailed instructions for application use, installation, and troubleshooting, improving user experience and system functionality.• Skillfully resolved IT-related challenges, including password and multi-factor authentication resets, and provided software installation assistance.• Assisted library patrons with navigating and utilizing resources efficiently and managing study space reservations, enhancing overall library operations.

Faridabad, Haryana, India

• Served as a full-stack developer to migrate a legacy application to Universal Windows Platform (UWP) desktop application. Tasked with re-engineering Account and General Ledger modules end-to-end using C#, LINQ middle tier, and MSSQL backend. Conducted API unit tests with Postman and managed Git versioning. Achieved 100% feature parity, reduced bugs by 30%, and cut rollback instances by 20%.• Mentored junior developers in transitioning to UWP development using PRISM architecture. Designed a learning track and conducted weekly evaluations, reporting progress directly to the manager. Accelerated junior developers’ onboarding by 30%, resulting in increased team productivity and faster feature releases.• Identified inconsistent code quality and security vulnerabilities during code reviews. Tasked to establish a code quality benchmark and monitoring system. Integrated SonarQube into the CI/CD pipeline to automatically assess code quality metrics and security risks. Reduced code smells by 60%, improved code coverage by 20%, and eliminated all critical security vulnerabilities, fostering a more robust and maintainable codebase.

Hyderabad Area, India

• Collaborated in capital markets of a top Canadian Bank; implemented ETL for external availability CSV files; programmed middle-tier with C# & LINQ, ensuring efficient ETL process per dynamically provided business rules. Integrated SSIS packages and wrote Stored Procedures for daily archival, reporting, and scheduling activities; ensuring ACID compliance for data integrity.• Collaborated on developing the GUI for the Locates’ booking process and Locates Manager tool in XAML, WPF using PRISM architecture. Specialized in file upload/read and real-time trade fulfillment through manual counterparty selection logic for pending approvals, achieving significant improvements in trade booking accuracy and streamlining the entire locates management process.• Implemented a waterfall approach to manage the manual selection of multiple counterparties in pending approvals, establishing a hierarchical decision-making process that effectively prevented negative inventory occurrences and enhanced operational efficiency.• Achieved 100% feature parity in the settings tab, encompassing crucial functionalities such as CRUD operations for new counter parties and their mailing lists, dynamic business rule configurations for booking, and more. Furthermore, significantly enhanced the report tab by re-engineering RDLC report generation, leading to a 2x increase in report generation speed.• Utilized on-demand cache calls with horizontal and vertical filtering techniques to optimize trade search efficiency in the GUI, achieving a 5x times faster page load time and significantly reducing trade search cycle time.• Co-created a Windows service to monitor CRUD with SQLTableDependency; developed a novel JSON distribution logic; monitored specific columns; scheduled off-peak bulk data loads; retroactive daily uploads; reduced congestion in IBM MQ & achieved data consistency between databases from 69% to 98%.

Masters Of Computer Science, Computer Science, 4.22/4.3

Thesis Title: Structural Embedding of Constituency Trees in the Attention-Based Model for Machine Comprehension Advisor: Dr. Vlado Keselj.

Bachelor’S Degree, Computer Science, 81.85%

Coursework Excerpt: Foundations of Algorithms, Data Structures, Operating Systems, Computer Architecture, Software Engineering, Computer.

High School, Secondary Education