There was no maintenance window for becoming myself

Content note: workplace harassment, assault, and transphobia.

Four years into my career, in 2019, I came out as a trans woman while working in California state government. I was transitioning in early adulthood, living in downtown Sacramento, and building systems for public service. I believed that useful work, honesty about who I was, and a clear request for what I needed would be enough.

It was not. In 2019, the harassment and assault became unbearable. I am not going to recount the assault here. Working in government did not protect me, and I learned in my body how wide the distance can be between an institution's stated values and a person's daily conditions.

That year, I asked for an all-gender bathroom until I felt ready to use the women's restroom. The request was denied. Instead, I was told to leave the building and walk two blocks down P Street, through heavy traffic, to use a bathroom. The institution answered a basic need by telling me to remove myself from the workplace to meet it.

In 2020, when the pandemic lockdown sent everyone home, remote work gave me separation and control that the accommodation process had not. The emergency was frightening and isolating. It was also the only reason I could keep working for the state. I survived more than anyone should have asked of me.

Government was always personal to me

I was born and raised in Sacramento and spent most of my life there, close to people working across many parts of government. Government was never an abstraction. It ran through the lives of people I cared about: capable people trying to make imperfect institutions work. That proximity drew me toward public technology because I saw both what government makes possible and how much its systems depend on the judgment and care of the people inside them.

I lived in downtown Sacramento until 2021, then moved to San Francisco, Albany, and eventually Davis when my fiancée began her PhD program at UC Davis. Each place gave me community; each also has a progressive reputation. But a progressive address is not an accommodation, and a stated value does not create safety by itself. Someone has to notice, decide, and own the remedy.

Coming out was not one event that ended in 2019. I still face hardship in my career because I am trans, even when it is quieter than what happened then. Now, in mid-career, with a director title and a record of delivering complex government systems, I still calculate whether a room is safe, how much to disclose, and whether speaking plainly will cost more than silence. Seniority gives me authority, not immunity.

That response taught me the difference between nominal availability and actual access. A restroom existed, but reaching it transferred the time, traffic exposure, and burden to me. The institution could mark the need as answered without changing the conditions that made the request necessary.

That distinction follows me into every transformation effort. I look for the point where a promise becomes a decision, funded work, an accountable owner, and a way for the affected person to know what happened.

There is no maintenance window for becoming yourself

Engineers know that the hardest modernization happens under live traffic. Energy reporting, workforce services, and Medicaid oversight cannot stop while the systems beneath them change. My transition happened under load too. There was no maintenance window for becoming myself. I still had to work, pay rent, sustain relationships, and plan a future while reexamining the assumptions governing my name, body, safety, and place in the world.

For years, I had been taught to question my own reality. Transition required me to reverse the direction of scrutiny. Instead of asking why I could not fit a given system, I began asking what was wrong with a system that required me to disappear in order to fit it. That reversal now sits at the center of my user-centered design practice. When a person cannot complete a form, understand a decision, use an interface with assistive technology, or remain safe in a workplace, their friction is evidence about the design. It is not proof that the person is defective.

Transition also made revision a habit. I learned to examine how I moved through the world and ask what aligned, what merely protected me, what caused harm, and what could become more honest, precise, or kind. Remaking myself was not a campaign to become acceptable; it was the work of aligning with what I knew to be true. I bring that stance to engineering and leadership, where processes, architectures, and beliefs can all be revised.

The analogy needs care. I was not a legacy system waiting for an upgrade, and the ways I survived before coming out were not technical debt. Much of what I built to stay alive deserves gratitude. Good modernization begins with comparable respect: understand what the current architecture protected, preserve the continuity people still need, and remove the constraints that no longer serve them.

That is where transition and public-systems modernization meet for me. Both ask what must remain stable, what needs to change, which inherited defaults are masquerading as reality, and how to move without abandoning the people living through the migration. Real transformation reaches beneath the interface to change the schema, ownership, incentives, and paths to redress.

Remote work is equity infrastructure

Remote work is often discussed as a perk or preference. For me, it became a condition of safety and continued employment. It reduced my exposure to unsafe people and spaces, gave me control over my immediate environment, moved more work into visible written channels, and let me spend energy on engineering instead of threat assessment.

Those conditions matter for many trans people and women, including trans women. Workplace equity includes commuting safety, bathroom access, scrutiny of appearance and tone, caregiving, disability, and who receives credit. Distributed work can reduce some of those pressures and make contributions easier to trace. It can also reproduce inequity if cameras become compulsory, access to leaders depends on proximity, or remote people are passed over.

Remote work becomes equitable when the operating system around it is deliberate:

Evaluate outcomes instead of rewarding physical visibility.
Document decisions and make asynchronous participation real, not ceremonial.
Treat accommodations as ordinary and confidential, not as a courtroom where someone has to prove enough pain.
Audit pay, consequential assignments, promotion, and retention by work arrangement.
Give in-person and remote staff equivalent ways to build trust, exercise influence, and lead.

The seams are where systems tell the truth

Public technology transformation is full of seams: policy to requirements, data collection to decision, vendor to agency, state to federal, interface to person. My transition made me unusually attentive to them. I want to know who owns the exception, what happens when the ordinary route fails, and who has the authority to repair harm.

At the Energy Commission, I architected and shipped an Amazon Web Services (AWS) Data Submission Portal for power plant and petroleum station energy reporting, supporting California's statewide demand forecasts and Integrated Energy Policy Report. The work required lineage across submitters, validation rules, pipelines, analysts, and policy decisions. In one system, a regulatory record could be traced end to end; in the workplace, an accommodation request had no comparable owner or accountable outcome.

At the California Public Utilities Commission, I directed the statewide Transportation Carrier Portal, an online licensing service for passenger carriers. A carrier needs to apply, correct a record, and receive a decision; it does not care which unit owns each field. Users experience the seam, not the organization chart. Internal complexity does not reduce the consequence for the person waiting on the other side.

I experienced institutional failure as a chain of handoffs in which each person could point elsewhere while I carried the consequence. Public systems fail people in the same shape. An eligibility rule can be correct in policy and impossible to complete online. A bilingual service can let one language grow stale. A model can return an answer without its evidence. A team can publish values without giving anyone power to stop harm.

I saw another form of silent exclusion as engineering lead and principal engineer for MyCareer.NJ.gov, New Jersey's statewide workforce platform serving 1.7 million residents. I kept more than 3,800 state-approved training programs current, found pipeline defects suppressing valid programs, enforced English and Spanish parity, and moved credential publishing toward the Credential Transparency Description Language (CTDL), a shared standard for describing credentials. That work made programs comparable and machine-readable rather than trapped in one system. A program can exist in a database and still be absent from a resident's world.

This is why I build gates, audit trails, fallback paths, source-linked decisions, and explicit ownership into public systems, and why I am precise about what a system can prove and where people must still exercise judgment. Kindness without precision is a promise nobody can use. Precision without kindness can become efficient cruelty. Engineering leadership requires both.

Difference is information

My experience does not make me an expert in anyone else's life. It makes me quicker to notice what I do not know, slower to confuse my default with a universal one, and unwilling to call a barrier an edge case because it is unfamiliar to me. Respect for difference requires curiosity and humility: ask, listen, test what you understood, and let the answer change the design.

As a people leader, this is the core of psychological safety for me: the confidence that you can name a risk, challenge me, ask for help, or think differently without losing status or belonging. I do not want a team that mirrors my thought process. I want people whose experiences, disciplines, and reasoning expose what I cannot see, with enough trust that we can learn from the difference.

I sometimes think of psychological safety as organizational observability. An application without good telemetry can fail while every dashboard stays green. A team can do the same: delivery appears healthy while risk, confusion, and harm remain hidden. Dissent, near misses, and the sentence "I do not understand" are signals a leader should want early. People are not components or data points, but leaders do choose whether truth can surface in time to act.

The same logic shapes how I approach accessibility, user-centered design, equity, innovation, and technology. I ask whose task becomes impossible under the default, whose reality informed the requirements, who pays for the workaround, and what becomes possible when difference is treated as information instead of friction. The answers shape how a service behaves at scale.

In practice, I put machine-checkable accessibility failures into continuous integration (CI), the automated checks run with each code change. Checks for contrast, semantics, keyboard behavior, and language parity should fail close to the change that caused the problem. Those gates prevent regressions; they do not prove a service is accessible. Manual usability work, assistive-technology testing, and direct feedback remain essential. I use the same pattern in leadership: automate routine safeguards and name a person responsible for what still requires judgment.

Grit should not be a staffing model

Sheer grit kept me working. It carried me through days when doing my job and protecting myself felt like two full-time jobs. I am proud of that endurance, and I am angry that it was required. Both things are true.

People often praise resilience after a system has spent it. As an engineering director, I do not want resilience to be a prerequisite for belonging on my team. When grit becomes an operating model, the cost of poor structure falls on people with the least room to absorb it, while those pushed out disappear from the measurement.

Trauma-informed leadership does not mean lowering the quality bar. It means removing arbitrary suffering from the path to meeting it. In practice, I try to:

Respond while a concern is still small instead of rewarding people for enduring until it becomes a crisis.
Make ownership, expectations, and escalation paths explicit.
Keep remote and asynchronous participation on an equal path to influence and advancement.
Separate a high performance standard from conformity in personality, body, schedule, or communication style.
Build enough shared context and leadership depth that nobody has to rescue the system alone.

The teams I lead know I care about evidence, testable claims, production gates, plain language, and clear ownership. They also know that a person can have a hard week without their career becoming suspect. People do better work when they spend less energy pretending to be invulnerable.

Modernization has to reach the workplace

Today, I lead engineering organizations and delivery across state workforce and federal health modernization. For the CMS Medicaid and CHIP Data Collection Tools suite, my teams maintain serverless TypeScript and AWS applications that states and territories use to report data for programs covering tens of millions of beneficiaries. Shared CI, role-based access, schema discipline, audit trails, and repeatable releases let the portfolio change coherently. The same principle drives my work with Fast Healthcare Interoperability Resources (FHIR), a standard for exchanging health information: interoperability matters when meaning survives institutional boundaries, not merely when systems exchange bytes.

On MyCareer.NJ.gov, I also designed a shared artificial intelligence infrastructure layer that runs in production and proof-of-concept features awaiting a state pilot. The design treats restraint as a technical capability through input validation, circuit breakers, kill switches, grounding checks, adversarial testing, and refusal when evidence cannot support a claim. In public service, a helpful-sounding answer that invents a training program, price, or outcome is a failure, so the system must be able to stop when its evidence is insufficient.

The technologies change by domain, but the questions do not. Who can safely use the system? What can be traced and explained? Where does uncertainty remain? Who bears the cost when it fails, and who is accountable for making it right? These questions apply to the organization as well as the software.

The workplace is part of the public service. A team harmed by its working conditions cannot sustainably build humane services for other people. A cloud migration that leaves avoidable human barriers intact is an incomplete transformation.

That is why the retreat from remote work worries me. Not every job can be remote, and not everyone wants it. Equity requires enough flexibility for people to contribute safely, with equal access to trust, consequential work, promotion, and leadership. An office can be valuable without becoming the price of admission to influence.

I will not turn harassment or assault into a gift. The harm did not make me better; what I chose to build afterward is mine. I lead with precision because vague support failed me, and with kindness because I know how expensive it is to remain capable while afraid. I look for the seams where institutions abandon people. I keep doing public-interest engineering because I believe systems can be made worthy of those who depend on them.

Endurance should be honored. It should never be a requirement for belonging.