Inside Meta’s Engineering Culture: Insights from John Miles White
John Miles White, former Director of Engineering on PyTorch and Meta’s Machine Learning Systems Lab (MSL), recently shared candid reflections on the tech giant’s evolving engineering culture, career dynamics, and the broader Silicon Valley landscape. Having stepped away from Meta, his insider perspective sheds light on the complexities and contradictions faced by engineers navigating today’s big tech environment.
The Changing Landscape for Engineers at Meta and Silicon Valley
John paints a nuanced picture of Meta as a company well-run from a business perspective but increasingly challenging as a workplace for employees who do not hold significant equity. He distinguishes between the experience of senior staff, who are often stockholders, and other employees who rely mostly on cash compensation.
A key shift he highlights is the labor market dynamic: unlike earlier years when engineering talent was scarce, today there is a perceived oversupply of engineers—except in niche areas like frontier AI research. This surplus reduces employee leverage, leading companies to scale back perks, reduce employee voice, and adopt a less accommodating culture. John notes this is not unique to Meta but reflects a broader Silicon Valley trend.
The consequences? More stressful work environments, a greater tolerance among employees for layoffs, and subtle shifts in organizational decisions such as reorganizations and talent retention strategies. The priority has tilted more toward business efficiency and less toward employee satisfaction.
The Promotion-Driven Culture and Its Discontents
One of the most striking revelations is the intense focus on promotions as the primary driver of motivation within teams, particularly in AI infrastructure. John observed that nearly every engineer's foremost goal was securing a promotion, which overshadowed genuine engagement with the work itself. This contrasts sharply with earlier Meta days, where career growth was less of an obsession.
This "promotion rat race" led to dysfunctional behaviors: teams shipping features they didn’t believe in simply to meet promotion criteria, and engineers prioritizing short-term wins over building clean, maintainable systems. Paradoxically, the skills that genuinely benefit one’s long-term career became decoupled from the promotion process.
John stresses that while the promotion-focused culture was highly prevalent, some teams like PyTorch bucked the trend by fostering a love of engineering for its own sake, holding a higher bar for promotions, and emphasizing craftsmanship over quick career advancement. This approach attracted engineers passionate about their craft, even if it meant slower promotions and lower compensation.
Lessons from PyTorch: Quality and Culture over Quick Promotions
PyTorch’s reputation for rigor and high standards made it a unique haven within Meta. Engineers there were often under-leveled compared to their true capabilities, but the credibility of PyTorch’s high bar helped them receive recognition when moving externally.
John reflects on his own experience managing PyTorch, where the culture prioritized sustainable success and healthy environments over rapid promotions. Although compensation might have been lower by about 20%, many chose this path out of pride and a desire to do meaningful engineering work.
Early Career and Impactful Work on Experimentation Tools
Before PyTorch, John contributed significantly to Meta’s data and experimentation tools, particularly the development of "Deltoid," Meta’s A/B testing framework. This work had a profound impact on how Meta conducted experiments to improve product decisions.
He recounts a pivotal career moment when many senior engineers left, thrusting him into a leadership role that accelerated his growth. The experience underscored how unexpected opportunities often arise from organizational churn, a common theme in tech careers.
John also notes the trend of successful tech startups emerging from products originally developed inside big tech—like Airflow and Optimizely—and reflects on missed entrepreneurial opportunities.
Programming Languages and the Julia Story
Before Meta, John was deeply involved in the Julia programming language, which aimed to combine the ease of high-level languages like Python and R with the speed of low-level languages like C. Julia’s niche was to offer high performance without sacrificing expressiveness, addressing frustrations with R’s slow execution and Python’s limitations.
He explains why R’s design leads to performance issues—due to dynamic features like overridable operators and lazy evaluation—which add runtime overhead. Julia’s approach challenged the assumption that high-level languages must be slow, but despite its merits, Julia remains less mainstream than Python or R.
Reflections on Academia vs. Industry
John expresses frustration with the misconception among many grad students and postdocs that industry is a fallback option if academic careers don’t pan out. He shares anecdotes of highly educated candidates struggling with basic programming tasks during interviews, illustrating the gap between academic research and practical engineering skills.
He encourages a mindset that values industry roles as challenging and rewarding careers in their own right, not as safety nets.
The Importance of Statistical Rigor in Tech
John is passionate about statistics, emphasizing the field’s tension between rigorous mathematical theory and practical application. He recommends the works of Larry Wasserman and Peter Arino for their honesty and rigor, contrasting them with more cavalier approaches common in the field.
He illustrates how misunderstandings of statistics can lead to problematic business decisions, such as misinterpreting A/B test results or setting goals that incentivize shipping code destined for deletion—a practice he finds counterproductive.
Career Advice and Leadership Lessons
Looking back, John regrets not taking advantage of leadership office hours offered by his skip-level manager early in his career, recognizing it as a missed opportunity for growth and connection. He advises junior engineers to seize chances to engage with leaders, as this can be invaluable.
His biggest piece of advice to his younger self? Be more confident and ambitious. John believes many talented people underestimate their potential and avoid risk due to self-doubt. He encourages embracing bigger challenges with discipline and optimism.
Final Thoughts
John Miles White’s reflections provide a candid and insightful look into the realities of working at a major tech company during a time of rapid change. His observations about labor dynamics, promotion culture, engineering craftsmanship, and career growth resonate beyond Meta, offering lessons for engineers and managers throughout the tech industry.
For those navigating big tech careers or interested in the intersection of engineering culture and business, John’s experiences underscore the importance of agency, choosing the right team culture, and balancing ambition with integrity.
Thank you for reading! If you enjoyed these insights, stay tuned for more conversations with industry insiders and deep dives into the evolving world of technology and engineering culture.