Instagram's engineering team writes about scaling, Python and Django, infrastructure, and mobile and open-source tools.
HN lists it among respected engineering blogs, and commenters cite it for Python at scale, PgQ cache invalidation and Cassandra latency work.
Python is a wonderful language where the speed of development and maintainability is concerned.
Django is a great framework for getting your web application and/or APIs built. The library support is also phenomenal (looking at Django Rest Framework for API development)
It is also unlikely you will outgrow Python/Django stack as Instagram and Pintereset are doing fine. https://engineering.instagram.com/what-powers-instagram-hundreds-of-instances-dozens-of-technologies-adf2e22da2ad#.nq4iq5lms
Rarely does the stack matter to startup success, why not pick up Steve Blank's
The Startup Owner's Manual about building THE RIGHT thing? Building the wrong product is the number one killer of startups why not focus there?
Memory is typically the scaling bottleneck for Python workers for web servers. One way to cut back drastically on this is to load as much as possible in startup and then fork for request serving processes. The problem is that python refcounting causes a lot of copy-on-writes for data that's really just used as read-only data.
This change allows you to run more workers with less memory.
Instagram has had a few good posts about how they've approached this problem, here's another:
https://engineering.instagram.com/dismissing-python-garbage-collection-at-instagram-4dca40b29172
I have met some members of Instagram's engineering team at Python conferences. They also run a dedicated engineering blog (https://engineering.instagram.com).
Despite the fact that they use an old version of Django, and the product has not exactly changed drastically since it became popular, they don't seem jaded.
Sometimes it's fun to build your own version of something regardless of what exists in the rest of the world. I'm sure nuances will become more prominent in time.
There is also PgQ from the https://wiki.postgresql.org/wiki/SkyTools package, which is fantastic and mature. It's like Kafka inside your PostgreSQL. When I was at Instagram, we implemented a multi-region cache invalidation system using PgQ that took advantage of atomicity between PgQ and writes to the source-of-truth data table. http://engineering.instagram.com/posts/548723638608102/
Honestly I think Uber switching was more of a business decision as their employees had trouble figuring out how to properly use Postgres. Instagram definitely has different data needs, but is probably much bigger than Uber wrt data. They use Postgres successfully via Django no less!
https://engineering.instagram.com/sharding-ids-at-instagram-1cf5a71e5a5c
There's a mention of Facebook regularly doing this in the summary section of this instagram engineering post: http://engineering.instagram.com/posts/548723638608102/
EDIT: Here's more info: http://www.datacenterknowledge.com/archives/2014/09/15/facebook-turned-off-entire-data-center-to-test-resiliency/