Ankur Patel

A data collection and showcasing dashboard platform to help publishers to monetize from sites by suggesting placement or sizes. It uses a forecasting formula to optimize the site layout by comparing the best-performing layouts based on ad viewability, ad blockers, and ad fraud.

I handled the dashboard creation, building a responsive single-page application (SPA) UI architecture in AngularJS. I also used Angular 1 and TypeScript. I then created a custom single sign-on (SSO) authentication for our product suite and integrated a REST API with PHP in Angular code.

During this project, I managed a 3-member team for the UI and back end. I also created a 100% automated build environment using Jenkins, Git, and AWS CloudFormation and achieved zero downtime deployment for the production to AWS. My other responsibilities included data collection and AdBlock reasoning tracking, where I monitored ad fraud and Active View metrics—such as eligible, measurable, and viewable impressions, the percentage of measurable and viewable impressions, and the average viewable time.

A data platform that enabled advertisers to target users based on opinions and emotions, collecting all its audience data from its media properties on a single platform.

We aggregated audience data from all of News UK's media properties—including The Sun, The Times, and talkSPORT—onto one platform for advertisers. The platform has a lot of first-party data about its users' preferences, opinions, and emotions, which could drive a potential 45% increase in ad engagement. This benefits advertisers and agencies as it enables them to create a custom audience and view the type of content they read. The platform will be available for brands as a planning tool.

I created a responsive SPA UI architecture in React and performed hands-on React and Redux development. I then integrated the Okta API with React for SSO authentication and used Node.js to build a microservices architecture.

During this project, I managed a 5-member team for the UI and back end. I created a 100% automated build environment using Jenkins, Git, and AWS CloudFormation and achieved zero downtime deployment for the production to AWS using the blue/green deployment strategy.

A platform that provides digital content and releases multiple new content groups daily to its customers who update their content library. The application is used to download, search, and purchase content in an industry-standard format, so users can drag or open it in any creative application they choose.

Digital Juice library includes animations, music, graphics, sound effects, stock footage, After Effects projects, layered Photoshop files, motion design elements, and special effects clips. It also has training videos, a Q&A support area, and a community of members who provide tips and advice. The platform further offers creative pieces and parts that can be integrated with its own custom assets, resulting in an original, highly polished finished project.

I developed the UI architecture for its new module in Angular 6, which was the Facebook posts kind, mobile-first UI. I also migrated the existing module from Angular 2 to Angular 6, implemented an Angular CLI structure and dependency, and upgraded the platform to make it mobile-friendly. Finally, I supported the DevOps team in deploying the platform's Angular 6 version.

A platform where users can create dynamic ads using predefined templates and target their ads to multiple platforms, like DoubleClick for Publishers and EB Advertising.

I developed the platform UI in AngularJS and Bootstrap and built custom components using Angular, Polymer, and D3.js for charts. I created the map component using Google API, a responsive feed component, and a nested component, such as video-in-video and video-in-slideshow. I also developed the color selection and photo story components. Additionally, I created the back-end REST API with Node.js and performed test-driven development using Mocha for the entire platform and custom components.

This project aimed to identify the provider company's logo from a video and give data to customers. It aims to develop a web application with UI, back end, and ML system to identify logos in the videos. The system will analyze video footage and accurately detect logos on the screen. To achieve this goal, the project will utilize machine learning and computer vision techniques.

The first step is to collect a large dataset of videos containing various logos. It should have a diverse dataset containing different types of logos, such as those seen in advertisements, sports events, and TV shows. Hence, with client help, I gathered a few types of logos and others we had generated using the Python augmentation library.

I trained a deep learning model on the preprocessed dataset using techniques using convolutional neural networks (CNNs). That model was tested and evaluated with a different set of videos, and I have used evaluation metrics called precision, recall, and F1-score.

Further, I created SPA using Angular and back end using Python Flask API to analyze video with a trained model to get results and show in tabular format in UI.

I worked single-handedly on this project with an E2E architect deployed on a client environment.