Developed web applications with Django, React and PostgreSQL. Worked as the lead developer for many of these projects.

Built a Python web scraper that aggregated school curricula, subjects, and lessons from the program websites of all Canadian provinces, saving 3 months of manual work for a large-scale e-learning platform.

Maintained automated unit testing and cloud deployment using CI/CD pipelines with GitHub Actions, containerized services with Docker, and deployed projects on the AWS platform.

Developed dilexsearch.com as part of MSc. thesis; an Angular application with a Node.js Express API hosted on AWS. The visualization and design got outstanding feedback from research communities.

Instructed undergraduate course CS215 for three semesters, covering web design, development (HTML5, CSS3, JavaScript, PHP), and SQL database programming.

Conducted full-time research in user-centric interface design, usability, and information visualization, leading to the publication of 3 papers at top UX conferences.

Collected user interaction data of an angular app (Dilex) with JS, stored in MongoDB by web socket channel through a node API. This data was preprocessed with Python and statistical analysis was conducted using R.

Engineered responsive, mobile-first web applications with Angular, integrating NgRx Store and lazy-loaded modules for improved caching and app fluidity.

Developed a project for managing tenants and properties of a large real estate company with ASP.NET / C#.

My thesis was about learning academic search behaviour and helping searchers to improve their search experience.

In this study, we propose a new method that aggregates keywords across a search engine results page (SERP), linking them visually to their source results. We created interactive and static interfaces, conducting a lab study to gauge their impact on utility and perceived value. The results highlight the effectiveness of using interactive keywords and visualization to summarize and link search results. Interestingly, whether keywords are shown alongside each result or aggregated over the entire SERP has minimal impact, emphasizing the importance of their interactive use in revealing relationships among search results.

We've built an academic digital library search interface called Dilex, aiding users in managing search tasks across mobile and desktop devices. In a controlled study, we compared Dilex to a standard academic search interface, finding increased user engagement and interaction with search results and personalization features. Participants spent more time on search result pages and showed enhanced engagement during resumed tasks. These results demonstrate the effectiveness of Dilex's visualization features in supporting cross-session search tasks and user engagement. This study showcases how semi-automatic task management and visualization can facilitate seamless cross-device search experiences.

Complex academic search tasks often span multiple sessions and devices, posing challenges for task resumption. Effective support for resuming tasks on mobile devices can utilize downtime effectively, while desktop/laptop support enables seamless access to mobile work. Our paper introduces an academic search interface facilitating cross-session, cross-device search using visualizations for browsing past topics and faceted navigation for examining saved documents. This approach optimizes search continuity and accessibility across varied environments and devices.

This research investigates a comparison between two different approaches for classifying emails based on their categories. Naive Bayes and Hidden Markov Model (HMM), two machine learning algorithms, have been used to detect whether an email is important or spam. Various combinations of NLP techniques- stopwords removing, stemming, and lemmatizing have been tried on both algorithms to inspect the differences in accuracy and find the best method.

Along with classifying emails, this paper also describes the methodologies used for automatic meeting scheduling by an intelligent email assistant. Users who regularly send or receive messages for setting up meetings will greatly benefit from this system as it will classify their emails and schedule their meetings automatically.