Amazon

S3 is one of the oldest AWS services and the scale is hard to describe. My project touched request routing logic where even a small inefficiency compounds across billions of daily requests. Engineers who work on it have been there for years and know the system deeply.

EKS is infrastructure for other people's infrastructure. Knowing that enterprises run their production workloads on what you're building is a different kind of accountability. Bellevue office is smaller than main Seattle campus which worked in my favor for access to senior engineers.

SageMaker team has a stronger ML engineering culture than the typical Amazon environment. Got to work on distributed training infrastructure. Technically interesting in a way a lot of AWS teams aren't. Big summer intern class, Seattle is great.

Vancouver office is smaller than Seattle which actually works in your favor. More direct contact with senior engineers, less of feeling like one of hundreds. The AWS work is the same quality and same scope regardless of which office you're in. Vancouver in the summer is genuinely great.

EC2 is a different scale than anything else. My project touched instance lifecycle scheduling and the number of hosts you're reasoning about is hard to internalize at first. Seattle intern community is massive, Amazon hires hundreds of SDE interns each summer so you're not short on people to connect with.

The scale is unlike anything else. AWS Lambda internals, actual hard problems. Seattle summer is great and the intern community across the company is massive.

AWS SDK work is more interesting than it sounds. You're writing tooling that millions of developers interact with, so even a small change has broad reach. Four month co-op term gave enough time to actually finish something. Vancouver office is relaxed compared to what people say about Seattle.