I led a team for a uni group project to consult, architect and implement an internal VetMed tool which stored placement information for students, and managed connections with placement providers.

The main bulk of this task was figuring out how to migrate all the existing data, which was a long PDF-based table, and convert this into a consistent schema.

After seeing a comedy skit from The Onion, I was motivated to create my own version - a real time voting app where event managers can pander to their audience.

Since my cloud experience was minimal at this point, I used Firebase, which simplified any complexity in dealing with cloud resources.

The main challenge this project faced was managing sessions. Since I wanted to avoid a massive cloud bill (even with 0 MAUs), I had to think about how to close down polling sessions when a time period expired. This was solved client-side using time calculations, but this comes with severe limitations.

The other challenge was managing concurrent users and maintaining live synchronisation. I used Firebase Realtime Database and transactions to handle this, which works for a limited number of users, but does not scale well.

This is a project I'm very interested in rearchitecting and reimplementing, using services like AWS API Gateway, CloudWatch EventBridge and DynamoDB Streams to build a more scalable and robust backend (even if I still have 0 MAUs).

I worked part-time to develop a MVP for a LLM-based project. The end goal was a pipeline that converted YouTube links or audio files into flashcards. Unfortunately, this was only a MVP and never made to market.

The idea was to use OpenAI Whisper and ChatGPT APIs to process these jobs, and this was before the AI bubble, so AI technology was still relatively new (at least for me).

I was mainly in charge of the frontend, and advising on architecture for the backend, but the MVP included a MERN stack application, which made API calls to run containerised jobs that used the OSS Whisper and made ChatGPT API calls

Given the ubiquitous number of AI companies and product services, I don't believe this product could be competitive nowadays (unless done really well). Nevertheless, this experience was insightful when it comes to developing applications and learning about UX.

This role was my first professional role in the software development space. During this time I worked on 2 projects and definitely grew the most as a software dev. It should be noted that I used AWS for the entirety of this role:

• I independently conducted a full migration from a manual deployment process to a full CI/CD pipeline - cutting deployment time by 63%, reducing human error by automating processes, and allowing developers to focus on the product rather than infrastructure.

  
  
  Skills:

  CloudFormation

  Infrastructure as Code

  Python

  CodePipeline

  Route 53

  CloudFront

  CodeBuild
  
  

• I provided technical mentorship to an intern, leading discussions on test coverage and fostering autonomy in learning and problem solving.

  
  
  Skills:

  Leadership

  Communication

  Collaboration

  Initiative
  
  

• I architected and developed an end-to-end AI object detection pipeline using MLaaS and the YOLO ML models for image processing. This included creating a batch processing system with event driven architecture for high elasticity, and sped up analysis over the client's manual process by 500%.

  
  
  Skills:

  Amazon SageMaker

  Event-driven Architecture

  SQS

  DynamoDB

  React

  Python

  AWS Lambda

  CloudFormation

  EventBridge

  CloudWatch

  Problem solving

  Solutions Architecture
  
  

• I overhauled and refactored a real-time data visualisation control centre dashboard to synchronise video, telemetry and data streaming from Kinesis Video Streams and DynamoDB, massively increasing reliability, improving performance, and allowing for extensibility and customisation through UI panels.

  
  
  Skills:

  Advanced React

  TypeScript

  Kinesis Video Streams

  DynamoDB

  S3

  Leaflet Maps

  Lightweight-Charts

  Communication

  Collaboration

  UI/UX Design