Personal
AWS
Creating EC2 Auto Scaling Groups for Spot Instances
Spot instances can be a cost-effective way for new cloud users like me who are trying to stay in the AWS free-tier or spend little money as possible when self-learning. It provides up to 90% discount off of standard EC2 instances. It's also great for dispensable or trivial workloads.
First, I created an Application Load Balancer with a new Security Group in order to distribute incoming http/https requests to instances. Because it's possible for my instances to become unhealthy or terminated, I needed to create a launch template for my Auto Scaling Group. I used Amazon Linux 2 AMI for my machines.
Next, I included the mixed instances policy, which allows for both on-demand instances (non-committal) and spot instances to be mixed together in my ASG. Because I am aiming for high-availability and a diverse pool of instances, I used the same VPC from my ALB for the subnets in my network.
Finally, I configured my scaling policy- I set desired and max capacity to 10. I set my minimum capacity to 5. I also set the default target value of 50% CPU utilization. So if my ASG determines that my CPUs are running below or above this threshold, it can properly accommodate for it. After playing around with my instances (terminating/stopping some), I can truly see the power of cloud and how they are automatically rebooted to desired capacity.
Using Amazon Polly and Transcribe for users with disabilities
Amazon Polly and Transcribe are two features I believe are very powerful for users with disabilities. For example, a visually-impaired or blind user is able to read computer screens but can use Polly to hear audio output of a text blog/post. A deaf user is still able to watch videos because Transcribe uses automatic speech recognition (ASR) for text scripts. A real-time feature is even available.
Polly
I first launched a Unix instance with a Wordpress blueprint. Then, I connected via SSH connect and set up a static IP in case I have to stop and reboot the instance later. I also performed a manual snapshot backup in case I needed to reboot an identical instance of the same configuration. To ensure access, I added a new IAM user access with Amazon Polly policy. After installing the AWS for Wordpress plug-in, blog posts are now audio enabled.
Check out the website here: http://3.233.3.182/
Transcribe
Additionally, these audio translations are available for download and can be stored into a S3 or LightSail bucket for transcription.
I took one of my favorite graduation speeches and uploaded it. I created a job in AWS Transcribe. It was completed in a few minutes. To ensure privacy protection, I enabled redaction measures to protection users' PII (personal identifiable information).
LightSail with Storage
Being beginner-friendly, I found LightSail very useful to learning more about the AWS cloud.
I decided to create a backup of media files hosted on the wordpress website created from my Amazon Polly project. Installing the WP Offload Media Lite plugin, I'm able to view any media file I upload in a blog post easily in an S3 bucket.
Shaped App
During the summer of 2020, I had the opportunity to collaborate with seven other college students from Cornell University and the University of Michigan to create a Christian faith and discussion focused app. As one of the product designers, I was responsible for the design and interaction of the app.
We were able to collect over 1000+ user survey response from Asian-American Christians globally from Facebook community groups. Followed up and conducted 30+ user interviews to inform our design decisions.