Started at Venn-Kassel today. Third position in four years. The intake form asked for “reason for leaving previous employer.” I wrote: “Refused to sign a fraudulent maintenance inspection.” The HR coordinator read it, looked at me, and typed something different. I didn’t ask what.

Assigned to Dock Section 11. Refit and overhaul on mid-range cargo vessels. My bench is at the back, near the coolant recirculation units. The hum from the pipes is 60 hertz. There is a bearing fault in pump 3 — I can hear it cycling every 8.2 seconds. A clicking sound underneath the hum. I mentioned it to the shift supervisor. He said he couldn’t hear anything.

The tool locker they gave me has a broken latch and the interior surface is sticky. I will use my own tools.

Shower at 0500. 42 degrees. 7 minutes. Tool case. Bench. Work. The order matters.

My mother’s tools are in the bottom drawer of the case. Ratchet set, micro-torque drivers, the plasma cutter with the worn grip that fits my hand because it fit hers first. She worked deep-range haulers for thirty years. Reactor maintenance. When she retired, she gave me the case and said, “Don’t let anyone else touch them.”

Nobody has. I check every night. Open the case, verify the count, close the case. Sixty-two items. Sixty-two every time. It doesn’t change. I know it doesn’t change. I check anyway.

The new bench has a good overhead light. The previous occupant left grease on the work surface that I spent forty minutes removing. The solvent they stock here isn’t strong enough so I used my own.

Dock Section 11 has nine other engineers. The shift supervisor is Hallam. He introduced himself by saying, “We run a tight schedule here. Deadlines matter.” I said, “I know.” He waited for me to say something else. I didn’t.

They brought in a Kessler-class hauler for engine overhaul. The work order said to replace the primary coolant seals and recalibrate the thrust governors. Estimated time: six hours.

I read the specifications first. I always read the specifications. The seals they’ve ordered are Venn-Kassel standard — compatible but not factory. The factory seals have a tighter compression ratio: 0.003 millimetres. At operating temperature, that’s the difference between a seal that holds for eighteen months and one that holds for eleven.

I told Hallam. He said the standard seals are what’s in stock and the client isn’t paying for factory. I said the specification sheet recommends factory. He said the specification sheet is a recommendation, not a requirement.

I replaced the seals. Standard, not factory. The work order was clear and I followed it.

The seal will hold. For eleven months. Then someone will need to replace it again. That person might be me. It probably won’t.

The canteen at Venn-Kassel serves meals at fixed times. Breakfast 0600-0630. Lunch 1200-1230. I eat at the start of each window because the tables are emptier. By 0620 the room fills up and the noise level makes it difficult to think.

The engineers from Dock Section 8 sit together. The ones from Section 11 scatter in twos and threes. I sit alone. I have been sitting alone for thirty-one days. I don’t know the rule. At some places, new people get invited. At other places, you’re supposed to approach. I don’t know which kind this is and there is no way to find out without doing the wrong one.

At Harlan Refit, an engineer named Priya sat with me every day without asking. She put her tray down and started eating. I didn’t understand this for three weeks. Then I understood it was the best thing that happened to me at that job. They transferred her after six weeks.

Breakfast: porridge, toast, juice. Every day. The tray has a small dent in the upper right corner. I use the same tray because it’s always in the same position in the rack.

The burns on my hands come from different years. The one across my right knuckles is from my first job — a coolant line that wasn’t properly bled before I opened the junction. My fault. I should have checked the pressure gauge a second time.

The ones on my left palm are from Harlan. A reactor housing that was hotter than the thermal reading indicated because the sensor was miscalibrated. Not my fault, but my hands.

The small round scar between my thumb and forefinger is from my mother’s ship. I was fourteen. She was showing me how to replace a plasma igniter and I grabbed the housing before it had cooled. She didn’t say anything. She put burn gel on it, wrapped it, and said, “Now you know what hot looks like before you touch it.”

I know what hot looks like. I also know what a seal feels like when it’s seated correctly — a specific resistance in the wrench at the final quarter-turn. You can feel it through the handle. Some engineers go by the torque reading on the meter. I go by the feel first and the meter second. The meter confirms what my hands already know.

Today I replaced fourteen seals. All seated. All confirmed. I checked each one three times.

Section 11 had a safety inspection today. The auditor walked the dock floor, checked the work logs, sampled three active jobs. One of them was mine — a thrust governor recalibration on a Pell-class shuttle.

He looked at my work log. It was five pages longer than anyone else’s. I record every step, every measurement, every tool used, every deviation from specification and the reason for it. He asked why my logs were so detailed.

I said, “Because the next person who works on this ship needs to know exactly what I did.”

He looked at me for a moment. Then he moved on.

After the inspection, Hallam told me the auditor had flagged my logs as “excessive documentation.” Not a violation. A note. He said I should consider streamlining. I asked what he wanted me to remove. He said, “Just make them look like everyone else’s.”

I didn’t change anything. The logs are correct. I don’t know how to make something correct look more like something less correct.

A mid-range freighter came in with a reactor fault. Reactor 4, port side. The crew reported intermittent power drops during transit — 3-5% loss, lasting 10-15 seconds, no pattern. Fleet sent them here for diagnosis.

I spent four hours with the reactor. Opened the housing. Checked the magnetic containment coils, the fuel injector array, the coolant circulation paths, the sensor grid. Everything read within tolerance.

Then I put my hand on the housing. Not on the instrument panel. On the reactor itself.

There’s a vibration that healthy reactors have. A steady hum, smooth, consistent. This one had a catch in it. Every ninety seconds, a tiny skip — so small the sensors wouldn’t flag it because it was within the noise floor. But I could feel it in my palm.

I traced it to a hairline fracture in the number three containment coil support bracket. A crack two centimetres long, invisible on visual inspection, just barely detectable on ultrasonic if you knew exactly where to point the probe. Under sustained load, the bracket flexes at the fracture, the coil shifts 0.2 millimetres, containment field wobbles, power drops.

I wrote it up. Hallam looked at the report and asked how I found it. I said, “I felt it.” He didn’t write that in the work order. He wrote, “Detected via ultrasonic inspection.”

On Fridays, the Section 11 engineers go to a bar on Level 4 of the station. I know this because they discuss it at 1600 every Friday. They haven’t invited me. This is the ninth Friday.

I don’t drink. I don’t enjoy bars. The noise and the light and the unpredictability of when someone will talk to you. I wouldn’t go if they asked.

But they didn’t ask. That is a different thing from not wanting to go. I know the difference. I have always known the difference.

Tool case: 62. Shower: 42 degrees, 7 minutes. Read: Harken-series reactor manual, chapters 4-6. Sleep: 2140.

Woke at 0200. Checked the tool case. 62. I know. I checked anyway.

New engineer in Section 11. Yolanda Ruiz. She’s young — maybe twenty-five. She talks while she works. Not to anyone in particular. To the ship, to the tools, to herself. She narrates what she’s doing: “Okay, number three bolt, you’re being difficult. Come on. There we go.”

This should annoy me. It doesn’t.

She asked me about the burn scars today. Most people look at them and don’t ask. She said, “Those look like reactor burns. How many years?”

“Thirty.”

“Thirty years of reactor work. That’s a lot of ships.”

“Yes.”

She went back to her bench. Twenty minutes later she brought me a coffee. I don’t drink coffee but I held the cup because it was warm and the dock floor is cold and the warmth was useful.

Hallam asked me to sign off on a refit today. Kessler-class hauler, full engine overhaul. Mettler did the work. I reviewed his logs.

The thrust governors are recalibrated to within tolerance — but barely. The lower end. Another 200 hours of operation and they’ll drift out. The coolant seals are standard, not factory. The reactor housing bolts are torqued to minimum spec, not the recommended median.

Everything is within tolerance. Everything will pass inspection. The ship will fly. It will probably be fine.

But “probably” is doing a lot of work in that sentence.

I told Hallam I wouldn’t sign off unless Mettler retorqued the housing bolts to median spec. Hallam said the ship is due for release at 1800 and there isn’t time. I said, “Then it releases at 2000.”

He said, “Koss, the work meets spec.”

I said, “I know. I won’t sign it.”

He stared at me. I waited. He’s done this before — the silence that’s supposed to make you reconsider. It doesn’t work on me. Silence is just silence.

Mettler retorqued the bolts. The ship released at 1945. Nobody thanked me. The bolts are at median spec. That’s enough.

Quarterly performance review. Hallam and the Section Manager, a woman named Deng I’ve spoken to twice.

Technical performance: “Exemplary. Koss’s work is consistently the highest quality in Section 11. Zero rework rate. Zero audit flags. Client satisfaction scores are the highest in the dock.”

Interpersonal performance: “Koss struggles to integrate with the team dynamic. Communication style is often perceived as blunt or dismissive. Reluctance to adapt documentation standards to team norms creates friction. Recommend: interpersonal skills development workshop.”

I asked what “interpersonal skills development workshop” involved. Deng said it was a two-day seminar on communication styles and team collaboration.

I asked if my work was the problem. She said no. I asked if my documentation was inaccurate. She said no. I asked what specifically I should change about how I communicate.

She said, “It’s not about what you say. It’s about how you say it.”

I have heard this sentence four times in my career. Each time I ask what it means. Each time no one can explain it in a way I can act on. They describe a thing I’m supposed to do differently, but they can’t describe the thing itself. It’s like being told to change the colour of a sound.

The first time I heard it I went home and practised saying the same sentence in different ways in my berth. I recorded myself. I played it back. They all sounded the same to me.

Things I count:

Tool case items: 62. Steps from berth to dock floor: 347. Bolts on a standard Kessler thrust assembly: 48. Seconds in shower: 420. Seals replaced since starting at Venn-Kassel: 211. Days since someone talked to me about something that wasn’t work: 14. Times Hallam has sighed during a conversation with me this week: 4. Engineers in Section 11 who make eye contact with me in the corridor: 2.

I don’t know when the counting started. It has always been there, the way the hum in the pipes has always been there. It isn’t something I do. It’s something that happens.

The counting is not a problem. Everyone says it as though it is — “you don’t need to check again” — but they are wrong. The checking is how I know. If I count the bolts, I know none are missing. If I count the days, I know where I stand. If I count the seals, I know exactly how much of my work is still holding somewhere in the dark.

My mother counted reactor cycles. I asked her once why she counted out loud. She said she didn’t know she was doing it.

I found the termination letter from my first job today. It was in the tool case, in the compartment under the micro-torque drivers. I don’t remember putting it there. I must have, but I don’t remember.

“...while your technical contributions are valued, the decision has been made to terminate your employment effective immediately. This action follows the incident of [date], in which you made allegations regarding the safety of the Series 7 reactor redesign that were determined to be unfounded by the review board...”

The Series 7 killed a technician eight months later. Coolant breach in exactly the junction I’d identified. His name was Perrin. I looked it up after. He was thirty-one. The review board issued a revised safety bulletin. Nobody contacted me.

I kept the letter. I don’t know why. I’ve been fired twice more since then and I don’t have those letters. Just this one. Maybe because this was the first time I learned that being right about something doesn’t protect you from the consequences of saying it.

I put it back in the compartment. Under the micro-torque drivers. I closed the case.

Sixty-two items. Plus one letter.

Day off. I don’t have plans because I don’t make plans on days off. I don’t know what other people do on their days off. Yolanda mentioned something about a garden level on the upper ring but I’m not sure if it was an invitation or just information.

I went to Dock Section 11. It’s empty on Saturdays. The lights are on standby — half brightness, 40% output. The dock floor is silent except for the coolant pipes. On workdays, conversation and tool noise sit on top of the 60-hertz hum and I have to sort them out. On Saturdays the hum is alone and I can hear it properly. I sat at my bench and listened to it for a while. Pump 3 is still clicking. 8.2 seconds.

I sat at my bench for a while. Ran my hands across the work surface. It’s clean — I keep it clean — but there’s a scratch in the coating near the left edge that I haven’t been able to buff out. It bothers me every time I see it. It doesn’t affect anything. It just bothers me.

I checked the Reactor 4 repair from last month. Pulled up the diagnostic logs remotely. The ship filed its latest report from a transit station eleven days out. Reactor running at 99.7% efficiency. No power drops. The bracket is holding.

99.7%. Not 100%. There’s always a margin. I built that margin as tight as I could and the number is 99.7 and that’s correct and it’s enough and I still wish it was higher.

I went back to my berth. Checked the tool case. Sixty-two.

Hallam called me into his office at 1530. Deng was there. The door was closed, which it usually isn’t.

He said they were restructuring Section 11. Reducing headcount. My position was being eliminated. It wasn’t personal. It wasn’t about my work. My work was excellent. The decision was based on “team fit and operational harmony.”

I asked what “operational harmony” meant. Deng said it meant the section needed engineers who could work collaboratively within the existing team culture. I asked if my work quality was the reason. She said no. I asked if there had been complaints about my work. She said no.

I said, “You’re firing me because of how I am.”

Neither of them said anything for a moment. Then Deng said, “We’re restructuring.”

I went back to my bench. Cleaned my tools. One at a time. In order. Put them in the case. Counted. 62. Closed it.

Yolanda’s bench was empty. Lower deck shift. I stood next to it for a while. Her tools were out — she leaves them out, which I would never do, but they were arranged neatly. She’d labelled her drawers in blue marker. The handwriting was uneven.

I thought about writing a note. I opened my mouth to say the words out loud, to test if they sounded right, the way I practise. “Thank you for the coffee.” That sounded correct. “Thank you for asking about the burns.” Correct. “Thank you for talking to your tools, because it made the dock floor sound like someone was home.”

I closed my mouth. I didn’t write the note. I don’t know if those are things you write down or things you’re supposed to have said when you had the chance.

Packed my berth. Two changes of clothes. The case. A datapad with my certifications. 14 minutes.

Three positions. Three terminations. The work was always rated excellent. My last three performance reviews all say “exemplary.” The exit paperwork says “restructuring.” The truth is in between, in a place I can’t reach because I can’t see it.

Kovac has independent repair work. No team. No harmony requirement. Just ships that need someone who will check the seal three times because three times is correct and I will not apologise for being correct.

Checked the case. 62. Plus one letter. Pump 3 was still clicking as I walked out. 8.2 seconds. Nobody will fix it. I wanted to fix it. I’m not allowed to fix it anymore.

347 steps to the dock exit. I counted them one last time.