- Wrote the code for the entire MVP using a stack React, TypeScript, NodeJS, and PostgreSQL- Architected a financial planning engine that translates a user's financial assets, liabilities, income, expenses, and goals along with all the financial rules applicable to them and government benefits (CPP/QPP, OAS, GIS, etc...) and tax credits (HST/GST, old age, etc...) into a larger set of linear inequalities. - Deployed backend using Serverless to AWS Lambda's for basic requests and use SNS events to launch more time-intensive work which then leverages WebSockets to communicate changes back to the front-end.- Front-end is deployed to an AWS S3 bucket and served through CloudFront- Designed the front-end for the website and the app which proved to be challenging given the amount of data and its complexity. The user has a lot of inputs that they need to provide and it is challenging to make it an easy and simple experience.- Created mobile app using Capacitor to wrap the existing React implementation and deploy it to the Google Play Store and Apple App Store- Learned to use the Plaid API to ingest user financial data and map it to our data structures to maintain support for adding assets manually without needing to connect a bank account- Implemented SEO into the React single-page app using Prerender.io and a couple of AWS Lambdas to redirect traffic to the cache.- Set up social media channels and our blog as our initial go-to-market strategy. The blogs are created using a custom CRM platform that we built which also allows us to change our SEO values on the fly.- Implemented an email distribution system using AWS Simple Email Service to outreach to users at specific times during the app's lifecycle. Can also send emails using our custom CRM platform.- Implemented Stripe as our payment solution

- I was a member of the community platform team which oversaw building the web application for On Deck fellows. Some of the main features it provided were the ability to search for other fellows and connect with them, the possibility to participate in and schedule online classes to learn more about starting a business or advancing your career, and a vast library of recorded content from past classes to watch. The platform is built in React + Typescript for the front-end, and GraphQL + Prisma + Typescript + PostgreSQL for the back-end, all while being deployed using Vercel. On Deck’s infrastructure also leverages a lot of no-code tools so it was necessary to build some interoperability between those various systems.- I helped deploy a new revision to the main dashboard for the community platform which aimed to address some usability concerns raised by some users. This involved taking the Figma documents provided by the design team and translating them into a usable and functional UI in React.- I worked on the first On Deck Onboarding Day (ODOD) which was a new initiative designed to have all new programs onboard their fellows on the same day to replicate the experience one might have at a university. This involved updating a lot of the onboarding flows to not only make this possible but also streamline and automate the experience so that we increased the number of users that made it through the whole process without intervention.- I worked on a proposal to rework the access control mechanisms in the app. The idea was to use a Resource Based Access Control (RBAC) system to allow content to be shared between programs as it was currently siloed to individual programs. The proposal went over how to achieve this by outlining everything from the database changes, to building a management app for program teams to easily modify who had access to which content, to how to fetch all visible content for a user in various situations.

- Helped start the Magnet Automate product which is a tool to help automate the ingestion of forensic evidence. When an agency obtains phones, computers, hard drives, or anything with some form of memory or storage, they connect them to their ingestion machines with their forensics software and run a large array of tools against them. This created a lot of inefficiencies as they needed people to start each software and they couldn’t just have someone monitor if the previous tool was done which created a lot of downtimes. Automate is built using React, Clojure, SQLite, and deployed it to both AWS and on-premises servers at the various agencies. It allows them to define their own workflows using a visual graph. Each node could be any CLI tool of their choosing but they had to configure it in a way that we understood how the input arguments worked and where it would output its result. The system was built in a distributed fashion, and they could add as many worker nodes as they wanted to increase the level of parallelization. The main server acted as a web server and the worker nodes simply had a light service that accepted commands from the master to execute the tools and report its progress.- Worked on the Magnet Examine product which is a tool for forensic examiners to review the evidence obtained from other products. Built in C# and WPF, my job was to help build the views used on the forensic evidence to explore, categorize, filter, tag, etc… I started on the project when it was a prototype and saw it through its release and a few of the initial versions. One of the biggest challenges of the app was the volume of data that needed to be displayed by the app which meant that we needed to optimize the database queries and virtualize a lot of the views for the app to remain responsive.- Participated in different agile frameworks on different teams including Sprint and Kanban- Leverage Jira for managing our tickets- Helped mentor various coops throughout the years