Over the last 18 months, there has been a massive changing of the guard at CU Direct in the Engineering Department, and there has been a mass exodus of many of the core players to match.
The *old* paradigm of CU Direct was a deep stable of onshore, full time, multi-talented software engineers that could all develop across any stack. In my tenure at the company, I saw more than a few engineers leave for FAANG companies. There was a clear focus on software excellence over speed of development, to a fault. The wide toolkits of the engineers allowed them to be extremely modular across teams and initiatives.
The *new* paradigm of CU Direct is a shallow core of onshore, full time software engineers (and some contractors), supplemented heavily by highly specialized and objectively much weaker offshore developers. These offshore developers outnumber the core at least 10:1, and what that means is that the onshore developers are often left to clean up code, or, in a worst case, merge shoddy code due to deadlines. In many scenarios, there is no onshore oversight on offshore code. This is something that is having a compounding effect across the software at the company, and we're starting to see stability and performance issues driven by this approach. To make matters worse, management has discussed this as a major pain point with engineering, when the obvious culprit (and only outlier) here is the paradigm shift, which stings for the onshore crew trying to hold things together.
Due to the fact that offshore has become the majority of the developers at the company, there is a very real expectation that meetings can occur any time between the hours of 7am and 9pm, and there is a heavy stigma associated with suggesting that you may not be available in non-standard work hours. Many teams have adopted a 7am start time to deal with this, which is essentially working a shift-workers schedule, but there is still an expectation that you are available during normal work hours (9-5) for onshore meetings.
As I mentioned previously, there is a significant disconnect between the main arteries at the company. Design is given different goals than Product, who is given different goals than Engineering, who is given different goals than Architecture, and all of these goals are redefined or changed on a monthly basis. This is devastating to software development, where clear through lines must be established and maintained for high quality development to take place. I'd guess a lot of this has to do with many of the middle managers leaving across the above departments, and the remaining managers simply being spread too thin for efficient conversation. The biggest stress point for the above disconnects seemed to fall in the lap of the PSEs and Directors.
Largely, most technical decisions at the company are made at the cadence of the Sales Team, with zero flexibility, and with little voice for the implications. Something might be significantly better if 20% more time was spent on it, or a new technology was introduced, but the shortcut will generally be taken to race to the next deadline or milestone. Down the road, there will be an attempt to pivot to the initial technical recommendation when what was promised (a subpar experience) is delivered.
To wrap it up, a couple small issues with larger implications, rapid fire: We seem to be a little listless in terms of our third party integrations. For servers we went from on-prem, to Pivotal, to Azure. For trackers, we went from Jira, to Pivotal Tracker, and then half back to Azure. Employees aren't allowed to use their own hardware, but the standard hardware can barely run some of the solutions that we're asked to run. DevOps has complete control over Azure (Engineering is completely blocked out), but they don't actually provide much support when it comes to deploying, setting up a ci/cd, or much of anything... they functionally exist as a layer of bureaucracy. Someone in technical management should be advocating for these issues.
