Learning on the Job 📚

You will be dropped into a project, tool, codebase, or dataset you don’t understand. That’s normal. Here’s how to get up to speed without panicking.

The 30-minute rule

If you’ve been stuck on the same problem for 30–45 minutes, ask for help. Your time isn’t free to the team — an hour of you spinning is worse than 5 minutes of a senior’s time.

But before you ask, do these three things (usually takes 15 min):

Search the internal docs / wiki / Slack history. 80% of “new” questions have been answered before.
Search externally (Google, Stack Overflow, official docs). Read the first 3 results properly.
Try to phrase the question clearly — often that alone unblocks you.

How to ask a good technical (or non-technical) question

Bad: “How do I do X?”

Good:

I’m trying to do X because [goal]. I’ve tried A and B. A gave me [error/output], B didn’t work because [reason]. My best guess is I need to do C — does that sound right, or am I missing something?

That kind of question gets you a great answer in 2 minutes. Vague questions get vague answers (or silence).

Reading unfamiliar work (code / reports / cases)

You don’t need to understand everything. Aim for:

What is this thing for? (the purpose)
Where does it start? (entry point / summary / intro)
What are the 3–5 key pieces? (main files / sections / stakeholders)
What don’t I understand yet? (write it down — ask later)

Don’t try to read the whole codebase / archive on day 3. You’ll drown.

Using AI tools responsibly

Most workplaces now allow AI assistants (Copilot, ChatGPT, Claude). Some rules:

✅ Use AI to explain concepts, generate first drafts, unblock syntax
✅ Verify anything AI produces before you ship it
❌ Don’t paste confidential company data into public AI tools — check your org’s policy
❌ Don’t submit AI output you can’t explain — you will be asked about it