Amazon

My first 7 years at Amazon, starting in Kindle and moving into the Amazon Publishing org, were incredible. My projects were varied and unique, and I was allowed to quickly expand my role from design-focused QA to animation, art direction, and print production while working with incredible teams.

During my last two years, the teams became stretched very thin, and project deadlines got much tighter with regular resource constriction in both budget and headcount.

The doc-writing culture that I had found incredible value in during past years became more toxic and seemed to be more valued than the work itself.

After COVID allowed most to WFH for multiple years with proven improvements in everything from productivity, quality of life, project efficiency, and communication, leadership decided that everyone would be better off in-office, with effectively no data to support the claim. In past years it had been made clear that all business decisions were made by leadership with data support, and the best that was offered was a "belief that office culture would improve team cohesion and productivity".

The introduction of LLMs on the HR side became nightmarish and remains so. Then came edicts to continually innovate processes and tasks using LLMs without updated resource documentation, reduction in access to technical knowledge holders, constricted deadlines, untrained teams with critical project workflow function, and with tools introduced in Q1 being quietly deprecated by Q3 or Q4.

The company has regular continual cycles of mass layoffs, salaries are typically not in line with other major tech companies, frugality is king, and frankly the leadership doesn't support their managers.

Do not let non-technical managers decide on technical project feasibility or risk, likewise with non-domain-specific projects managers have only oblique knowledge on, especially in periods of resource constriction. It will severely constrain your teams to unattainable goals, put them in the position to make unrealistic requirements, and you will lose critical functionality.

Stop building the ship while sailing; those 'hacky' fixes will become the established tools, and team members will not have time to go back to revise or fix bugs.

Reduce outsourcing to teams with rotating heads requiring regular retraining for critical functionality.