Keisha Williams became a track maintenance supervisor for the New York City subway system seven years ago, right around the time she stopped believing in the phrase "hundred-year flood." She's 31 now, managing a crew of eighteen workers responsible for keeping trains running through a drainage system designed when her great-grandparents were children. We met at a diner in Queens on a Tuesday afternoon, between the morning rush repairs and whatever the evening would bring. She ordered coffee and barely touched it.
This conversation is a composite based on reporting about urban transit infrastructure workers, though Keisha herself exists primarily in the space between what's documented and what's lived.
How do you explain your job to people at parties?
Keisha: I say I'm in transit maintenance and pray they leave it there. They never do. Someone always asks what that means, and then I'm explaining that I supervise track crews for the MTA, and then they want to know what that actually means, and eventually I end up describing how I spend most of my time trying to keep hundred-year-old drainage systems working during storms they were never designed to handle.
That's usually when people get this look. Like they've just realized something they'd rather not think about.
The technical version: my crews maintain track infrastructure. The drainage systems, the pump rooms, everything that keeps water from turning subway tunnels into underground rivers. We've got this whole network of drains and pumps rated for 1.75 inches of rain per hour.1 Which would be great if storms still respected those numbers. Instead we get 2.5, 3 inches in an hour, and I'm coordinating emergency responses instead of doing actual maintenance.
When did you realize this wasn't the job you thought you were signing up for?
Keisha: July 2021. Three years in, still thought I was building a real career in infrastructure engineering. Then we got that storm. Not even a hurricane, just a summer thunderstorm that decided to park over the city and empty itself.
I watched the Lexington Avenue line flood in real-time on the security feeds. Water pouring down the stairs like someone opened a fire hydrant. People trapped on trains. I'm on the phone with my crews trying to figure out which pump rooms to prioritize when everything's underwater at once.
We got the trains running again. We always do. But that was the night I realized I wasn't maintaining infrastructure anymore. I was performing triage on a patient that was already dying.
I started keeping a different count after that. Not just work orders completed, but how many times per month we're responding to flooding that shouldn't be physically possible given the historical rainfall data. The math is pretty straightforward. The system was built to handle 1904 weather patterns. We're living in 2025. That gap used to be theoretical.
What does a typical day look like now?
Keisha: When I started, maybe 20% emergency response, 80% scheduled maintenance and actual long-term projects. Now it's flipped. We're constantly in reactive mode.
I've got crews that were supposed to be replacing aging track components. Instead they're clearing debris from drains after storms, testing pumps that failed overnight, inspecting flood damage. This past July we had multiple flooding events that forced us to skip stops, reroute trains, shut down whole sections.1 Each time that happens, there's the immediate crisis, and then there's the aftermath nobody sees. Corroded electrical components that'll fail in three months instead of three years. Track bed erosion that accelerates everything. It compounds.
My morning routine: check the weather forecast before email. If there's rain predicted, I'm already mapping which sections are most vulnerable, which crews I'll need to reposition, which pump rooms need extra attention. I've got this whole color-coded system in my head. Red zones flood if we get more than 2 inches in an hour. Orange zones can handle 2.5 if the pumps are all working. Yellow zones are probably fine unless multiple things go wrong.
The "probably" is doing a lot of work in that sentence.
The MTA announced $700 million for flood protections. Does that change anything?
Keisha: [Laughs] You want the diplomatic answer or the real one?
Diplomatic: it's a significant investment in climate resilience and we're grateful for the resources to fortify the system.1
Real answer: $700 million sounds like a lot until you understand the scale. We've got 665 miles of track, 472 stations, pump rooms and drainage infrastructure spread across the entire system. That money will help with flood barriers, improved drainage, upgraded pumps. But we're not rebuilding from scratch. We're retrofitting a Model T to handle modern highway speeds. You can upgrade the engine, improve the brakes, but you're still working with a frame designed for a completely different world.
The other thing nobody mentions is timeline. That's 2025-2029 capital funding. We're installing protections against flooding that's happening right now, today, while we're sitting here. Your house is actively on fire and someone hands you a check to install a sprinkler system that'll be operational in four years. Great, thanks, but what do I do about tonight?
What's it like psychologically, maintaining infrastructure you know is inadequate?
Keisha: Some days I feel like I'm playing elaborate pretend. We fix things, we improve things, we implement new protocols, and everyone acts like we're solving the problem.
We're not solving it. We're managing it. There's a difference.
I've got this one pump room in Lower Manhattan. I won't tell you which line because I don't need that on record, but the pumps in there are older than I am. We maintain them, we keep them running. Every time I'm down there I'm thinking: these were designed for a climate that doesn't exist anymore. The rainfall intensity, the frequency, the duration. All changed. And I'm standing there with a clipboard making sure the 40-year-old pump is properly maintained according to specifications written before "climate change" was even a phrase people used.
The weird part is that I'm good at this job. Really good. Two promotions in seven years, which doesn't happen often. And there's this uncomfortable thing where I know I'm advancing professionally because climate change is making my expertise more valuable. The worse things get, the more the organization needs people who can keep the system limping along.
That's not a comfortable thing to acknowledge at 31.
Do you talk about this with your crews?
Keisha: Depends on the person. Some of the older guys, the ones who've been doing this for 25, 30 years, they've got gallows humor about it. "We're not maintenance workers, we're hospice care for infrastructure." That kind of thing.
The younger workers are different. They don't have a before-times reference point. For them, this is the job. Constant emergency response, climate adaptation, managing failure. They're incredibly competent at crisis management because they've never known anything else. But I worry about burnout. The infrastructure isn't sustainable, obviously. But the people maintaining it? That's a whole other question.
I had a crew member quit last month. Smart guy, good worker, just done. He told me he couldn't keep pouring his energy into a system everyone knows will need fundamental reconstruction eventually. Went to work for one of those climate adaptation consulting firms. Gets paid more, doesn't have to wade through flooded tunnels.
I didn't blame him.
What keeps you doing it?
Keisha: [Long pause] Honestly? I don't have a great answer.
Inertia, partly. This is what I know. But also the trains have to run. That's not philosophical, that's just reality. Eight million people in this city depend on the subway. If my crews don't show up, if we don't do the work, people can't get to their jobs, to hospitals, to their lives.
So we show up.
There's also this thing where I'm weirdly optimistic about the long-term while being completely cynical about the day-to-day. Does that make sense? I genuinely believe we'll figure out climate-resilient transit infrastructure eventually. We have to. But I also know that tomorrow I'll be dealing with pump failures and flooded track sections because it's supposed to rain.
My girlfriend says I've developed very specific compartmentalization. I can talk about the systemic inadequacy of century-old drainage infrastructure and then immediately pivot to discussing which vendors have the fastest turnaround on replacement pump parts. You have to hold both things in your head or you'd lose your mind.
What do you wish people understood about urban infrastructure and climate change?
Keisha: That it's all connected. Everything. When the subway floods, the water backs up into the sewer system, which is also over capacity. That affects water treatment. Storm drains overflow into streets, which affects road infrastructure. Power systems fail, which affects everything else.2
People think about climate change as this future thing, this abstract environmental issue. But it's infrastructure. The systems that make modern urban life possible, all designed for a climate that doesn't exist anymore. And we're all just making it work somehow. Patching, upgrading, adapting, responding.
The other thing: there's no cavalry coming. No one's going to swoop in and fix this. It's going to be people like me and my crews, showing up every day, keeping things running with duct tape and institutional knowledge and increasingly inadequate drainage capacity.
That's climate adaptation in practice. It's not inspiring, it's not heroic, it's just work. Endless, exhausting, necessary work.
Do you think about leaving? Doing something else?
Keisha: All the time. I've got a civil engineering degree. I could go work on new construction, green infrastructure, climate-resilient design. All the stuff I thought I'd be doing when I graduated. Build things instead of keeping old things barely functional.
But then I think about who'd replace me. Who'd know which pump rooms are most critical, which track sections flood first, which crews can handle the worst situations. It's not ego. I'm not irreplaceable. But there's accumulated knowledge that matters. And if everyone with that knowledge leaves for better opportunities, what happens to the system?
So I stay. For now. I tell myself it's temporary, that I'm building experience I'll use later on better projects. Maybe that's true. Or maybe I'll be here in twenty years, still watching weather forecasts and managing flood response, watching the gap between system capacity and climate reality get wider every year.
Either way, the trains have to run tomorrow.
So I'll be there.