It is possible to talk even to a runtime engineer, but you need to be on their enterprise plan and spend millions per year with AWS. It may sound brutal, but AWS prioritizes based on the account size.
Given the AWS Active mention, the author may be living off free AWS credits and has a modest spend afterwards.
Not sure if a seven-week investigation is the best use of engineering resources at early startups. Once you hit the limits of PaaS platforms, going EC2/Fargate/Kubernetes seems like a more pragmatic option. AWS Lambda is not for async jobs; once you return HTTP, it's done. Not convinced this is a bug.
However, he did manage to get to the front page of Hacker News, and the write-up is quite decent (minus the AWS blame part).