Built several full stack workflows for the intake of new users and management of current users including reusable components, Foundational API structures, and Unit and Integration tests.
Optimised several heavyload API workflows using parallelization, query optimization, and caching capabilities to improve the efficiency of data access successfully reducing load times by 50%+ and reducing the number of redundant API requests and network load
Spearheaded the project to rebuild and redesign the legacy software using React.js, Java Spring Boot, and MSSQL to improve data flows, user experience and customer satisfaction
Standardized demographic data across over 1 million data rows and 20 database tables to conform with ONC standards while ensuring cross compatibility with HL7 and MDS, hence reducing dirty data by 65%
Led and Managed a team of 5 developers as scrum master, holding us accountable to higher productivity and efficiency standards, and ensuring that we were consistently 1-2 sprints ahead of deadline.
SeroTracker PT Technical Team Lead, PT Frontend Developer
July 2021 - Present
Leading the design and implementation of 2 new cross pathogen dashboards for data visualization and data submission using Next.js, Flask and postgreSQL in a global data standardization effort
Worked on a major restructuring of the SarsCov-2 dashboard improving latency by over 90%, updating the code base with standard industry convention, connecting to several external APIs and abstracting several features in preparation to scale to a multi-pathogen dashboard smoothly and seamlessly
Led a team of developers and designers, managed workload distribution, and conducted performance tests on diversing system design decisions to ensure optimized solution adaption
Collaborated with the Data and Partnerships teams and other industry partners during the iterative building of the dashboards to ensure data privacy, security and feature satisfaction
Maintained and updated the existing SarsCov-2 dashboard striving to consistently improve the product
Projects
Peera: worked on building a course load management startup to help university students to better organize their time and find community in their classes. Built the reminders, team chats, mentorship and test prep schedule features. Used React.js, TailwindCSS, Flask, PostgreSQL and EWS
WantWords Reverse Dictionary: improved efficiency of the WantWords CNN on several NLP metrics by 2-5% by augmenting it with BERT perceptron to efficiently distribute the weightage of different score metrics. Done as part of a paper reproducibility and improvement project. Used PyTorch, Beautiful SOUP, and WordNet
Twitter Data Analysis: conducted data analytics study on vaccine and COVID related tweets. Used Twitter API to extract 1400+ tweets and cleaned, filtered, grouped, and annotated all tweets based on the task. Created TF-IDF and sentiment analysis charts for each group to find the topics of interest in these conversations.
McGill AI Podcast: Hosted an AI podcast hosting MILA researchers on topics like trends, accessibility, and ethics.
Pentago Twist AI: Created an agent with an Alpha-Beta pruning algorithm to play and win at Pentago Twist. Used game analytics and statistics to create a strong heuristic with a 90% win rate against human players.