I improved my understanding of serverless architecture, object storage, API design, workflow orchestration, AI service integration, IAM permission design, and observability on AWS. The project helped me connect individual AWS services into one practical system instead of learning them separately.
I learned to reduce an ambitious product idea into a realistic MVP. The original idea was a real-time AI communication assistant, but the final implementation focuses on post-conversation reflection because it is safer, cheaper, more reproducible, and technically defensible for a bootcamp project.
I proposed an original use case based on a real communication problem I observed: people often understand a topic partially but struggle to organize and defend ideas during conversations. I independently redesigned the scope, selected AWS services, and created a workshop structure that another learner can follow.
I followed a weekly worklog, documented implementation steps, and included testing, monitoring, security, cost, and cleanup sections. I also kept the project bounded to avoid unrealistic production-level complexity.
The project improved my ability to explain architecture choices clearly. I practiced describing why each AWS service is needed, what trade-offs were made, and how the system handles privacy, cost, and failure scenarios.
During the bootcamp, I learned from mentors, peer discussions, AWS Study Group materials, and event sessions. I used feedback to refine the idea from an overly broad AI assistant into a focused, deliverable cloud project.
The main design challenge was balancing ambition and feasibility. I solved this by moving from real-time processing to asynchronous processing, adding transcript upload as a fallback, and using Step Functions plus CloudWatch to make errors observable.
My personal contribution includes: