Why I left

A personal account of power, silence, and leadership — after six years at Meta

Having handed in my Meta badge and laptop just a few days ago, I’ve been reflecting on what actually broke me.

It wasn’t the amount of work — even if working in a different timezone while raising young children without a family support system isn’t a walk in the park.

It wasn’t disinterest in VR either — I’d giddily gotten the Oculus Rift DK2 in 2014, and later built with fascination for HoloLens.

And it wasn’t performance ratings — my lowest year-end rating at Meta was Exceeds Expectations.

And no, I wasn’t fired either.

Tracing the source of my first burnout forced me to go back in time: to the moment I accepted a job I hadn’t wanted in the first place.

The burnout came as a surprise, because I’d spent over two decades in Tech on four different continents without burning out. Academia, Silicon Valley startups, Microsoft, my own startup.

It took me a while to understand why this time was different, and looking back the story adds up to three core forces:

1. Power went unchecked, making the organization increasingly inefficient.

2. Speaking up became risky — silence was implicitly rewarded.

3. The kind of leadership I practice (empathy, multi-perspective thinking, psychological safety) became devalued.

This is how those forces unfolded.

---

Resistance

A Facebook recruiter called me about a job with Oculus, a VR headset maker Facebook had acquired a few years earlier, twelve days before the due date of my second child. A new job was the last thing on my mind. And Facebook in particular wasn’t attractive — this was right after the Cambridge Analytica scandal had broken.

We had moved from Silicon Valley to Singapore, planned to stay through our kids’ toddler years. I told the recruiter about my “condition,” expecting never to hear back. Pre-#MeToo, the rule for a woman in Silicon Valley was: “DO NOT TALK ABOUT YOUR FAMILY, EVER!”.

But the recruiter persisted.

Eight weeks later, newborn in tow, sleep deprived, recovering from surgery, pumping every four hours — I started interviewing. I was too exhausted to wear the polished mask I’d worn in Silicon Valley. I was simply myself.

To my surprise, I passed the interviews and found something unexpected: kindness. Empathy. From every single person. Little gestures that made me feel: These are good people.

So I pushed harder: I openly said I only wanted to work here if climbing the career ladder isn’t required — often another mortal sin in Corporate Tech. They still offered me the job.

For the first time in my career, I thought: I can belong here without hiding who I am.

We made the leap. My partner paused his tenure-track professorship; we moved our family across the ocean; I accepted the role of Engineering Manager for Oculus in Switzerland.

I still wasn’t sure about the parent company, Facebook. So during new-hire orientation in London, I grilled everyone about Cambridge Analytica. I expected denial. Instead, everyone said: “Yes, we screwed up” and “We are fixing it,” and provided concrete evidence (not sharing details here) that they meant it.

I concluded the negative media portrayal of Facebook was exaggerated.

---

Falling in love with the work

Early on, my manager asked me to support a prototyping team — a multi-disciplinary group who lived and breathed VR. At the time we were one of only two such teams in XR Tech. I was impressed by the influence our small, remote team had: we helped shift our org from a tech-first mindset (“we build it because we can”) to one centered around user-experience (“we build it because it adds value for users”).

In partnership with the Zurich tech teams we also developed early Mixed Reality prototypes, starting with Passthrough and progressively more advanced approaches to blending real and virtual worlds. We had reviews across and up the chain all the way to Boz and Mark Zuckerberg - ultimately shifting the direction of the Meta Quest line from “VR only” toward Mixed Reality as a core offering.

I concluded that such bottom-up influence was only possible with low-ego leadership. And I admired those leaders for it.

Having grown up with an abusive alcoholic father, I’d developed a strong early radar for ego, lust for power, and the inability to admit mistakes. What I saw at Meta in those years felt like the opposite: calm feedback, accountability, humility.

So I took initiative to go on four external speaking engagements every year — implicitly and sometimes explicitly making the case that Facebook was misjudged. I spoke at schools, universities, with policymakers in Berlin, and annually at our developer conference Connect (recording, 2023).

My first two years’ average performance rating: Greatly Exceeds Expectations.

I loved the work and I pushed hard.

---

Cracks in the Facade

Crack #1: Diversity Hiring and the Silent Warning

As a prototyping team representing the customer it was vital that our team composition at least somewhat represents the customer (vs. being a monoculture). While trying to recruit a more diverse team, I was unexpectedly contacted by Employment Law — an internal process no one talks about. I was questioned without being told the context. Only later did I learn that disciplinary action, even termination, could result if they found any violation.

I was eventually fully cleared.

But the experience sent a clear message: as the main breadwinner who had just moved her family across continents, I could not actively recruit for diversity the way I believed was necessary.

My first silencing.

---

Crack #2: Child Safety

Around 2022–2023, I became increasingly concerned about child safety in parts of the metaverse. In third-party apps and also in Horizon, I observed children and young teens using adult accounts and being exposed to interactions that, in my opinion, appeared age-inappropriate and raised serious concerns about child-safety for me. (Similar concerns have also been discussed publicly by some users, e.g. 1,2,3,4).

I raised these observations internally and tried to address them through the channels available to me. I worked with Policy partners and proposed a privacy-preserving technical approach to more reliable age estimation.

While the issue itself was broadly acknowledged and involved complex trade-offs, I experienced a growing tension between the urgency I personally felt around child protection and the rate of progress I saw. I also saw the framing shift - reviewers highlighted PR risk to increase the likelihood of internal traction. I was ultimately told that pursuing a technical solution wasn’t encouraged.

Meta later introduced industry-leading parental controls and much stricter age verification, which is laudable progress! Still, I couldn’t shake the conviction that these measures were arriving years later than I was comfortable with, and that we still weren’t reliable enough at catching kids impersonating adults — a sentiment echoed by some independent creators of 18+ Horizon Worlds who implemented early aspects of the technical approach I had originally proposed.

This was my second silencing. Not because nothing was being done — important measures were taken — but because of the disconnect between the bar for reliability and urgency I believed was required, versus what I saw.

I was told
“It took us years to get into this problem, and it will take a while to get out of it“

but to me this was a SEV-0 — the kind of issue where teams mobilize to fix immediately, not over years. That disconnect stayed with me.

Crack #3: Collapse of Psychological Safety

Then came the crack that hit my personal trauma buttons.

I began hearing senior colleagues describe negative experiences with an Executive — shifting moods, shutting people down, feedback that felt personal rather than focused on the work. I didn’t believe it at first, not at this company, and definitely not this person whom I so admired. Then I witnessed it.

In one meeting, he repeatedly directed strong language (“fucking stupid”, 5+ times) toward a senior PM, making it difficult for that PM to disagree or course-correct the conversation. And two new contingent workers who had just given a demo initially assumed the outburst was related to them.

I don’t mind strong critique. I do mind when power dynamics make disagreement unsafe.

My childhood fight reflex kicked in; I spoke up. Then, in the following days, I went from fight → flight → freeze. I couldn’t be in the same room with him.

I thought this was “my trauma.” Until a junior PM told me about a recent review with the same Executive:

“In my house we don’t shout. When he raises his voice, I freeze. And when he shouts he doesn’t listen, so all your prep is pointless.”

A mid-level data scientist said:

“No one dares to correct him when he’s wrong. Even the Directors just say ‘we’ll take an Action Item’”

This behavior contradicted our Community Engagement Expectations. Yet there was no safe way to report it. And once you were on this Executive’s shitlist, it was nearly impossible to regain influence.

My third silencing.

But the worst part was what it signaled: the system had become dysfunctional.

Organizations can survive bad code easier than suppressed truth. Human debt is harder to repay than tech debt.

---

The Cultural Shift — Inefficiency at Scale

Once I saw the pattern, I saw it everywhere.

Another Executive had once told me “It’s hard to be a jerk here.” That may have been true then, but around Covid time the culture shifted.

PSC, our semi-annual performance measurement process, appeared to lose important checks and balances. Where HR had once pushed back when a manager’s proposed performance rating for a person was controversial, the stance now seemed to become: “It’s your decision in the end.” I witnessed a senior case where communication issues were kept off the rating despite pushback from several managers, contradicting five years of PSC norms I had experienced.

In another PSC change, a person’s contributions to “Helping others,” called the People axis, appeared to become less important: Where subpar performance in this area once lowered ratings, this later had no impact — subtly shifting incentives away from fostering team spirit and toward individual advancement.

The no-feedback culture went all the way to the top. Before a demo to Mark Zuckerberg, handlers adjusted the room temperature to Mark’s preferences; in another, his and Boz’s favorite chairs were brought in. And once, when Mark offhandedly mentioned his contact lenses drying out during a VR demo — years into the headset’s production lifecycle — half the senior leadership on a 20-person call bent over backwards to investigate, as if product strategy had suddenly pivoted to eye care.

I realized senior leadership had become an echo chamber, not a culture that survives feedback.

These issues led to inefficiencies no dashboard measured — particularly acute on the EM ladder and more pronounced the higher up you went:

• Days spent crafting slides that wouldn’t trigger negative reactions rather than “tell it like it is”

• Weeks lost managing perceptions instead of solving problems.

• The “screaming chain”: escalating issues up, up, up instead of peer-to-peer collaboration.

• The increasing requirement for political maneuvering to secure scope.

• The gradual shift where loudness and dominance were rewarded over thoughtfulness.

The inefficiencies these dynamics created went unseen as they fell in patriarchal blindspots. Our Year of Efficiency overlooked the biggest inefficiency of all: ego.

My moral compass tells me to reduce ego-motivated maneuvering. Succeeding in this environment would have meant increasing it. I wasn’t willing to play. And thus I no longer belonged.

---

Fast Forward to 2025

When fact checking was deemphasised, I tested something: I posted the same intentionally inflammatory comment on Facebook and our internal message board called Workplace. It was removed on Workplace. It stayed up on Facebook. The contrast illuminated internal vs. external priorities for me: Inside Meta, removing offensive content was considered vital for maintaining social order. Outside of Meta, we didn’t seem to care.

Our Community Engagement Expectations group, CEE, intended to ensure a respectful working environment, increasingly felt like a tool to discourage internal dissent. Following my post, CEE called me and reminded me to “disagree & commit,” which I experienced as a stark warning.

Then came rhetoric about the company needing to be “more masculine.” Gender needn’t have been invoked to make the point — especially in an industry so heavily skewed towards one gender. And “more masculine” was interpreted by many as more internal competition, increased aggression, less empathy — precisely the cultural shift that had already eroded efficiency.

At this point, it was simply confirmation for what I’d known for a while: I didn’t belong in Big Tech any longer. And I didn’t want to.

---

So, What Actually Broke Me

Ultimately, it comes down to three things:

1. Unchecked power creates profound inefficiency.

The more senior someone becomes, the fewer people challenge them. Without checks, ego often fills the gap. Ego in turn poisons true partnership, and with it efficient (and joyful!) collaboration. The inefficiency this causes is not 10% — it is exponential. A company trying to be efficient must build structures that counteract human nature.

2. I cannot be silenced.

I am too seasoned and too anchored in my moral compass to wear a mask. And this kind of integrity isn’t incompatible with capitalism — it is a prerequisite for sustainable capitalism.

3. Meta made the kind of leadership I practice ineffective.

I lead with vulnerability, multiple perspectives, honest communication, and strong support for my team. Research and my track record showed these traits produce better outcomes. Yet they were increasingly interpreted as junior, weak, or incompatible with the “new culture.”

When diverse leadership styles are devalued, leaders like me cannot lead effectively.

---

This Generalizes

This isn’t just Meta.

We are watching a broader societal trend: people who choose power over integrity rise fastest. Ego is rewarded. Dissent is punished. The costs — human, societal, economic, and environmental — accumulate quietly until they explode. This is why I can’t be silent; at some point silence becomes compliance.

I don’t exempt myself from responsibility for having stayed as long as I did. I share this not to absolve myself, but because understanding how systems like this keep well-intentioned people compliant longer than they expect matters if we want to change them.

Listening to others helped me place my own experience in a broader context:

• Careless People by Sarah Wynn Williams (former Director, Global Policy)

• Brian Boland (former VP) testifying before the Senate committee for Homeland Security

• Kelly Stonelake’s substack (former Director, PMM)

• allegations by Jason Sattizahn (former User Researcher) around child safety in the Metaverse

• a story about power from Natalie Ponte (early Meta employee).

---

Today — And What Comes Next

This marks my retirement from Meta — and from tech in general, aligned with my plan to retire from tech in my 40s. This by the way is the reason I can speak more candidly now, because I no longer depend on my reputation in the industry.

Healthwise, I’m fortunately well on my way to a near full recovery. I’m now focused on three things:

1. Raising two boys to become caring, empathetic, resilient humans who know they are enough just as they are.

2. Building a mental health app aimed at helping to democratize access to recent advances in trauma treatment for adults with childhood trauma.

3. In the medium term, should the right opportunity arise, I might choose to give back to the European tech ecosystem.

In the end, the very forces that wore me down also made my path unmistakably clear. So long Meta.

It’s been a ride.

---

A Note of Gratitude

I had the privilege of working with exceptional engineers, PMs, designers, user researchers, artists, QA teams, and managers. Many are still at Meta, doing important work every day. This post is not a judgment of anyone who stays — I understand the economic and reputational constraints. It was an honor to work alongside you.

A particular thanks goes to my team, a source of deep joy especially but not only through our early years, and the reason I have held on for this long. You rock and it was an honor to support you.

Thank you also to my direct managers who supported me until the end - thank you.

Also, being human makes me naturally imperfect. So lastly a heartfelt apology to anyone who may have felt wronged by me – do reach out if you feel like it.

\