Of all the petroleum sub-disciplines, drilling is the one whose mistakes call you at 2 a.m. The drilling engineer argues with a rock that's pushing back in real time — and bills tens of thousands of dollars a day while the argument is unsettled.

There are two kinds of drilling engineer, and most of them are both at once: the one who plans the well at a desk, and the one who owns it while it's being drilled. The planning version has a calm, almost architectural day. The operations version has a phone that goes off at 2 a.m. The same person switches between the two depending on whether their well is on the rig this week.

So a day depends entirely on where the well is.

If the well is still on paper

A well in planning is a stack of interlocking design problems, and the drilling engineer works them in roughly this order.

First the trajectory — where the well goes and how it curves to get there. In a Permian or Appalachian development that means a horizontal that has to land in a target window a few feet thick, two miles down and two miles out, without colliding with the dozens of existing wells around it. This is anti-collision work, run in a directional planning tool like Landmark's COMPASS, and it is not optional: nudging a wellpath to stay a safe distance from an offset is a calculation you show your work on, because the failure mode is intersecting a live wellbore.

Then the casing design — the nested steel strings that line the hole. Each string has to survive burst, collapse, and tension loads, and the program has to set casing seats at depths that keep the open hole inside the mud-weight window: heavy enough to hold back pore pressure, light enough not to fracture the formation. Get the window wrong and you either take a kick or lose your drilling fluid into the rock. Tools like StressCheck and CasingSeat exist precisely because this is too important to do in a spreadsheet, though plenty of people still try.

Then torque and drag and hydraulics. As the wellbore gets longer and more horizontal, simply sliding pipe in and out — and rotating it — runs into friction that can exceed the rig's mechanical limits. You model it in WellPlan (Halliburton/Landmark) before you ever feel it on the rig floor. The hydraulics side asks a parallel question: can you circulate fast enough to clean the cuttings out of a two-mile lateral without the equivalent circulating density climbing past the fracture gradient? The mud is doing three jobs at once — pressure control, hole cleaning, cooling the bit — and the window where it does all three is narrow.

And underneath all of it, the well-control plan: kick tolerance, kill-sheet calculations, the arithmetic you hope you never need but rehearse anyway.

If the well is on the rig

Now the day is different. Now there's real-time data streaming off the rig over WITSML feeds — rate of penetration, weight on bit, torque, standpipe pressure, hookload, and the MWD/LWD logs coming up from just behind the bit — and the drilling engineer is watching it the way a flight controller watches telemetry. Often from an operations center hundreds of miles from the rig.

The morning ritual is the daily drilling report. Every hour of the last 24 is coded — drilling, tripping, circulating, and the one category everyone hates, NPT (non-productive time): the hours lost to a stuck bottom-hole assembly, a mud-motor failure, waiting on cement, weather. Those codes roll straight into cost, tracked line by line against the AFE — the authorization for expenditure, the budget the well is supposed to live within. A drilling engineer can tell you the burn rate of their rig to the hour, because the difference between a well that comes in under AFE and one that blows through it is often a single bad trip.

Then there's geosteering — the live collaboration with the geologist, nudging the bit up or down on the strength of LWD gamma and resistivity to stay in the pay zone as the formation rolls and faults in ways no plan fully predicted. The plan said the target was flat. The rock disagreed. You adjust.

The 2 a.m. call

Every drilling engineer has the stories, because operations don't keep office hours. The pipe gets stuck and you're running back-off and jar-force calculations while the rig crew waits. You take a kick — formation fluid entering the wellbore — and the well-control procedure you rehearsed on a spreadsheet is suddenly the only thing standing between a controlled kill and a very bad night. You lose returns into a thief zone and have to decide, fast, how to keep the hole full. This is the only petroleum sub-discipline where the mistakes are measured in hours and the worst ones are measured in lives, which is why the culture around it is so unforgiving about procedure.

The quiet inefficiency

For all the sophistication of the real-time stack, a striking amount of the drilling engineer's analytical day still happens off to the side. The integrated databases (Landmark's EDM, for instance) hold an enormous amount — every NPT event, bit run, mud property, and cost on every well — but the moment an engineer wants to ask a question across wells ("how did the last ten laterals in this area behave through this formation?"), the answer often gets assembled by hand, exported to Excel, and reasoned about there. The data is captured beautifully and queried clumsily. Offset-well learning, the single richest source of "don't make that mistake again," is frequently a manual archaeology project.

Where this connects to what we build

Drilling is the discipline where the gap between data captured and decision supported is widest — gigabytes of real-time and historical detail, and an engineer still pulling offset wells into a spreadsheet to see a pattern. That gap is the whole reason Workollab exists: not another giant platform, but practical tooling that turns the data a small operator already owns into something they can actually ask questions of, running on their own infrastructure, with the engineer's judgment kept firmly in the loop. The rig already generates the signal. The job is to stop losing it on the way to the decision.

The drilling engineer builds the path everything else flows through. They just deserve to learn from the last well without re-excavating it by hand.

Tools & terms named in this piece

COMPASS (anti-collision / trajectory), WellPlan, StressCheck, CasingSeat, EDM/OpenWells — Halliburton Landmark / Engineer's Desktop suite; WITSML real-time data standard; MWD/LWD; ECD (equivalent circulating density); mud-weight window and pore-pressure/fracture gradient; NPT and the daily drilling report (DDR/IADC); AFE; geosteering; well control (kick tolerance, kill sheet). Workflow and tool details drawn from current Landmark/Halliburton documentation and drilling-operations sources, 2025–2026.

One of Workollab's oil & gas engineering pieces.