If you talk to enough engineering leads, a pattern shows up in how they describe upskilling. A few years ago the answer to “how do we get better at this?” was usually a course, a certification track or a conference. Today more of them mention something smaller: a recurring half hour where someone works on one stubborn skill, gets a quick reaction and tries again. It is anecdotal rather than measured, so treat it as a trend worth watching, not a statistic. This piece looks at the rise of deliberate practice in technical teams, why it appeals, where it goes wrong and what it asks of leaders.
From big events to small reps
Courses and certifications are good at transferring a shared vocabulary. They are less good at fixing the thing you do badly at 4 p.m. on a Thursday. Many people finish a week-long programme feeling energised and return to a job where the same awkward step still takes them an hour.
The drill-based approach grew out of that gap. Instead of widening what you know, you narrow what you attempt: one command sequence, one troubleshooting move, one explanation to a stakeholder. You repeat it several times, compare your result with a standard and change one thing. It sounds almost too modest, which may be why it took so long to spread.
Why teams are drawn to it
Three reasons come up repeatedly in conversation.
- It fits real calendars. A fifteen-minute drill survives incident weeks that would cancel a full-day workshop.
- It targets actual weak spots. People pick the thing that slowed them last sprint, so the learning is relevant on Monday.
- It produces evidence. A drill has a time, a pass mark and a note about what changed, so managers can see progress without guessing.
There is also a cultural benefit. Drilling a weakness out loud normalises not being perfect. Teams that do this together often report that people ask for help earlier, though again this is observation, not data.
What the better versions look like
The teams that seem to get value share a few habits. First, a drill has one target, written as a sentence. Second, the time box is short and respected. Third, feedback arrives within the session, from a checklist, from a reviewer or from an automated check. Fourth, only one variable changes between attempts. Fifth, someone records a line or two about the outcome.
For data and machine learning groups, the targets tend to be concrete: re-running a training pipeline from a clean checkout, tracing a failing job to its cause, or describing a deployment step so a colleague can repeat it. A solid reference for orienting those drills is the sort of structured plan found in a stage-by-stage roadmap for data scientists and engineers working together; pick one stage and drill it rather than trying to read the whole map.
Where it goes wrong
Not every attempt works. We see four recurring failure patterns.
- Drills that become meetings. A fifteen-minute exercise grows a slide deck and an agenda, and the repetitions vanish.
- Safe targets. People choose things they already do well, which produces a warm feeling and no growth.
- Feedback without teeth. If the only response is “nice work”, nothing changes.
- No follow-through. One burst of enthusiasm, then nothing for two months.
Leaders can prevent most of these with small rules: a hard time limit, a requirement that every drill names a weakness, and a standing question at the end, “what will you change next time?”.
The role of outside eyes
A noticeable shift is that more teams are asking outsiders to watch a drill rather than deliver a lecture. The reasoning is straightforward. A coach observing ten minutes of real behaviour can name a habit that colleagues no longer notice. That is cheaper and quicker than a generic session, and it keeps the focus on the person’s actual workflow.
Where to find such people is a practical question. One option is DevOpsTrainer, a directory that lists DevOps trainers, instructors and coaches you can browse by area, including SRE, DevSecOps, MLOps and AIOps, then read the formats on offer for individual upskilling or team sessions. Whichever route you take, ask for a trial observation of a drill before committing to anything bigger.
A week in the life of one team
To make this concrete, picture a platform group of six. On Monday morning they pick the week’s weakness from a shared list: “reading a failed rollout without panicking”. Tuesday and Thursday, two people take ten minutes each in a sandbox, running a staged failure and narrating what they check first. A third person holds a checklist and ticks what was covered. After each round the narrator changes one habit, perhaps looking at recent events before logs. On Friday the group spends five minutes comparing notes and records a line.
Nothing about that week is dramatic. Over a quarter, though, the list of drilled weaknesses becomes a quiet record of what the team improved, and new joiners can run the same exercises on day three instead of absorbing habits through luck.
What leaders should do next
If the trend appeals, start small. Pick one team and one recurring weakness. Schedule a fifteen-minute drill twice a week for a month. Appoint someone to keep two lines of notes after each session. At the end, ask three questions: did the drilled task get faster, did mistakes in review decline, and did people want to continue? If the answers are mostly yes, widen it. If not, change the target before changing the format.
Skills in fast-moving areas make good first targets. A short read on the training that modern IT operations people tend to need can help you pick which capability deserves the first drill, then convert one item from it into a ten-minute exercise.
Caveats worth keeping in mind
Deliberate practice is not magic. It will not replace broad study when you meet a field for the first time, and it works best on skills with clear right and wrong outcomes. Soft skills such as negotiating or explaining a trade-off can be drilled too, but the feedback is fuzzier and needs a trustworthy reviewer. And the research on how far practice structure explains expertise remains debated, so claims of guaranteed results should raise eyebrows.
Still, the underlying idea has survived because it matches everyday experience. Everyone has watched a colleague repeat the same slow workaround for years. Everyone has also seen someone fix a stubborn habit in a week by isolating it. The trend is simply organisations noticing that second pattern and giving it a regular slot in the calendar.
Looking ahead
Expect more tooling around it: sandboxes that reset in seconds, scoring scripts that run after each attempt and libraries of small exercises owned by the team. Expect more pairing of drills with short written logs, since evidence is what turns a pleasant habit into a funded one. And expect the old advice, “just get more experience”, to sound a little thinner each year. Experience still matters. It matters most when someone is steering it at a specific weakness and checking whether the steering worked.