When you lead the technical side of a project, you still ship, but you're also the person people come to for reviews, design decisions, production issues and "quick questions" that rarely are.
If you don't plan for that, interruptions take the whole day and the hard technical work gets pushed into your evenings. This is the structure I use to keep room for both.
Where the Time Actually Goes
Start by being honest about where your time goes. For me it's some mix of:
| Activity | What it looks like |
|---|---|
| Deep work | Designing and building the hardest parts of a feature, refactoring, performance work |
| Unblocking others | Code reviews, answering questions, pairing, debugging with a teammate |
| Planning and design | Writing technical designs, estimating, reviewing requirements with product |
| Communication | Stand-ups, status updates, client or stakeholder discussions |
| Operations | Investigating production issues, on-call, release support |
| Growth | Learning new tools, reading, mentoring, improving team processes |
You can't drop any of these. What you can do is stop them from competing for the same hours.
Tip
Track your time loosely for one week: just a note every hour of what you actually did. Most engineers are surprised by how fragmented their days are, and that data makes the rest of this guide much easier to apply.
Step 1: Protect a Daily Deep-Work Block
Hard engineering work needs long stretches without interruption. Thirty minutes between two meetings is enough to answer a review. It isn't enough to work through a tricky concurrency bug or design an API.
Pick a two- to three-hour block, ideally when you think most clearly, and defend it:
- Put it on your calendar as a real event so people can't book over it.
- Close chat and email, or set your status to "focusing, back at 11:30".
- Decide the task the night before, so you don't spend the first 20 minutes choosing.
Mornings suit a lot of people because fewer meetings have started, but use whichever slot you can actually keep free.
Step 2: Batch Code Reviews and Questions
Reviews and questions are some of the most useful things you do for a team. One good review can save hours of rework. Answering each notification as it arrives, though, wrecks your focus.
I set two or three review windows a day, for example after stand-up, after lunch and late in the afternoon. People soon learn they'll hear back within a few hours, and that's fast enough for nearly everything.
Note
Make one exception: a teammate who is completely blocked, or a production issue. Agree with your team on how to flag those (a direct mention, a dedicated channel) so true emergencies still get through.
Step 3: Run Meetings Like Code
Meetings are fine when they have a clear purpose. Give them the same care you'd give a pull request:
- If there's no agenda, skip the meeting. When nothing needs deciding, an async update is enough.
- Put meetings next to each other so the rest of the day has usable blocks.
- Finish with decisions and owners. Write down who does what by when before people leave.
- If you're not needed, decline politely and ask for the notes.
Step 4: Plan the Whole Week
A daily to-do list keeps you busy. A weekly plan makes sure you work on the right things. At the start of each week, spend 20 minutes on three questions:
- What's the one most important outcome this week? A feature shipped, a design approved, a risky migration done.
- What could block it? Missing requirements, dependencies on another team, an unanswered question.
- Who on the team needs my help this week? A review that's waiting, a teammate stuck on a new codebase.
Then schedule the main outcome into your deep-work blocks first, before other people's priorities fill the week.
Step 5: Write Things Down
A lot of leading a project is making it less necessary for you to be in the room. Writing lets you do that:
- Design docs before big changes, so decisions get reviewed early and are remembered later.
- Short decision records ("we chose X over Y because…") in the repo or wiki.
- Runbooks for production issues you've fixed, so the next person can handle them without you.
- Pull request descriptions that explain why the change was made, as well as what it does.
Each question the docs answer is one less interruption for you.
Step 6: End the Day on Purpose
Use the last 15 minutes to set up tomorrow:
- Leave yourself a note about where exactly you stopped and what comes next.
- Reply to or clear anything that's blocking a teammate, so they can start tomorrow without you.
- Update the ticket or your status so nobody has to ask where things are.
- Pick tomorrow's deep-work task.
Then stop working. You'll do better over the years with steady, rested focus than with the odd late night.
A Sample Day
This is roughly how it can look. Adjust the times to your team and time zone:
| Time | Focus |
|---|---|
| 09:00 – 09:30 | Check messages, review the plan for the day |
| 09:30 – 09:45 | Stand-up |
| 09:45 – 10:15 | Review window 1: pull requests and questions |
| 10:15 – 12:30 | Deep work (notifications off) |
| 12:30 – 13:30 | Lunch, away from the screen |
| 13:30 – 14:00 | Review window 2 |
| 14:00 – 15:30 | Meetings, design discussions, pairing or mentoring |
| 15:30 – 17:00 | Second focus block or smaller tasks |
| 17:00 – 17:15 | Review window 3, end-of-day notes, plan tomorrow |
Warning
Don't expect a perfect schedule every day. Production incidents and urgent requests will happen. Treat the routine as the default you come back to, and don't feel guilty when a day breaks it.
Summary
Keep time for hard technical work, be available at predictable times, and write things down so the team doesn't need you for every answer. If you only change two things, start with one protected deep-work block and batched reviews.



