Upgrade & Secure Your Future with DevOps, SRE, DevSecOps, MLOps!

We spend hours on Instagram and YouTube and waste money on coffee and fast food, but won’t spend 30 minutes a day learning skills to boost our careers.
Master in DevOps, SRE, DevSecOps & MLOps!

Learn from Guru Rajesh Kumar and double your salary in just one year.

Get Started Now!

Start, Stop, Continue: A Retro for Your AI Career Skills

Start, Stop, Continue: A Retro for Your AI Career Skills
Artificial Intelligence

You shipped the model, the dashboard stopped flashing red, and the team went quiet for a day. That quiet is the perfect moment for a start-stop-continue retrospective. The practice is tiny, yet it pays off for anyone growing your own working skills in artificial intelligence and AIOps: you look back at a stretch of work and sort what you noticed into three lists. What should we start doing? What should we stop? What is working well enough that we should continue?

This guide is written for people who are early or mid-way in AI work and want their skills to grow on purpose rather than by accident. We will treat the retrospective as a personal and team tool, give you concrete examples for each list, and show how a few lines of honest writing can steer a whole year of learning.

Why a skills-focused start-stop-continue retrospective works

Most AI practitioners learn in bursts. A new framework appears, you binge tutorials for a weekend, and three weeks later you cannot remember what stuck. A retrospective inserts a pause between bursts. Instead of asking “what should I learn next?” in the abstract, you ask what the last stretch of real work taught you, which gaps got exposed, and which habits quietly saved the day.

The three-list shape matters. Start forces you to name something new. Stop forces you to let go, which is the part most people skip. Continue protects what is already working, so you do not throw away a good habit in the enthusiasm of change. Together they turn a fuzzy sense of “I should improve” into three small, checkable promises.

Start: habits worth adding after a release

Starts are the fun list, so keep them modest. A good start is something you could begin tomorrow without asking permission. Here are examples that tend to come up after a machine learning or AIOps release.

  • Start writing a one-paragraph note on every model change. Record the change, the reason, and the outcome you predict. Six months later this note is worth more than the commit message.
  • Start reading incident write-ups from other teams. Fifteen minutes a week gives you a library of failure patterns you have not yet met yourself.
  • Start pairing with someone from the operations side. Data scientists who sit beside the people paged at night learn quickly what “production ready” means.
  • Start keeping a small evaluation set that reflects real, messy inputs. It becomes your reality check when a flashy benchmark says otherwise.

Notice that none of these require a course or a certificate. They are practice habits. Hiring managers usually notice people who can account for odd behaviour in a system, and these habits build exactly that skill.

Stop: habits to retire gently

The stop list is where careers quietly improve. It asks you to look honestly at time spent on things that feel productive but are not. A few common ones in AI work:

  • Stop tuning a model for a week before checking the data. Most surprising results trace back to inputs, not architecture.
  • Stop collecting tutorials you never open. A saved list is not a skill.
  • Stop saying yes to every new tool trial. Each trial costs setup time; pick the ones tied to a real problem you have right now.
  • Stop leaving alerts on that nobody reads. They train the team to ignore the one that matters.

Write stops as things you will not do, and give each a replacement where possible. “Stop tuning first” becomes “check the inputs first.” A stop without a replacement tends to creep back within a month, because the underlying itch has nowhere else to go.

Continue: protect what already works

Do not skip this list. It is easy to treat “continue” as a courtesy, but it records the quiet routines that made a hard stretch survivable. Maybe your team always wrote a rollback plan before a deploy. Maybe somebody always checked the training data against a recent sample before retraining. Maybe you always blocked an hour on Friday to read logs without a particular goal.

Writing these down does two jobs. It shows you which of your habits are worth defending when schedules get tight, and it gives newcomers a written picture of how the team really works. For your own career, the continue list is a quiet evidence file: a log of steady, useful behaviour that backs up any conversation about how you have developed.

A worked example from a small AIOps team

The following situation is made up, but it will sound familiar. A small team had rolled out an anomaly-detection feature for infrastructure metrics. Two weeks after launch, they held a retrospective. Their starts: add a weekly review of false alarms with the operations engineer, and begin tagging each alert with the reason it was judged useful or noisy. Their stops: stop retraining on the same schedule regardless of drift, and stop treating accuracy as the only number worth reporting. Their continues: keep the plain-language release notes, and keep the habit of asking the on-call person for feedback within a day of any change.

Six lines. Yet the career effect for each person was large. The junior engineer learned how drift shows up in practice. The analyst learned to speak in the operations team’s language. The lead found that the habit most worth defending was something nobody had thought to call a “best practice” at all.

Turning the lists into a personal growth plan

If you want the retrospective to feed your career, add one more step: convert the starts into learning items. A start such as “keep an evaluation set of messy inputs” becomes a small project. A start such as “read other teams’ incident reports” becomes a recurring calendar slot. Keep the plan short, and review it after your next stretch of work.

It also helps to browse how people in neighbouring fields describe their own routes. A discovery site like BestPilotsNow, built around aviation with separate corners for people, training providers, learning material, practical resources, a discussion board and written pieces, is a handy reminder that every profession keeps its learning scattered across several kinds of places. Your AI path will too: some of it lives in documentation, some in conversations, some in write-ups, and some only in the logs of your own systems.

Connecting the retro to the wider AI picture

Retrospectives pair nicely with the reading and thinking you already do. If you want an example of how AI shows up in day-to-day decisions, this piece on how AI sharpens the choices businesses make gives context that is worth turning into a start item for your own list. For a different angle, the article on AI’s growing role in classrooms is a useful prompt for stops and starts around how you explain your own work to non-specialists.

Common questions people ask

How often should I do this?

After any natural boundary: a release, a sprint, a quarter, a long incident. Thirty minutes is plenty for one person. For a team, aim for an hour at most.

What if my stops are out of my control?

Split them into the ones you can act on today and the ones you can only raise. Put the first group on your list and bring the second to whoever owns the decision, with a concrete suggestion rather than a complaint.

Will this slow down the work?

Briefly. In return you avoid repeating avoidable mistakes, which is the slowest thing of all in a field that changes this fast.

Your next thirty minutes

Open a blank page. Write three headings. Look back over your last stretch of work and add two or three items under each. Choose one start to try this week, one stop to practise saying no to, and one continue to protect on purpose. Put a reminder on the calendar to look at the page again after your next release.

That is the whole practice. Build it into your routine, and your skills will grow in the directions your real work demands, which is where careers in AI and AIOps are actually made.

Subscribe
Notify of
guest
0 Comments
Oldest
Newest Most Voted
0
Would love your thoughts, please comment.x
()
x