How do I build my time estimation skills for project

Unfortunately, a lot of timeline crafting is trial-and-error. As an E3, you are going to mess this up a lot. In a vacuum, I recommend just blindly tacking on 50% to whatever initial deadline you come up, so 2 weeks would become 3 for example. I have mentored so many E3s and I have literally never seen them overestimate a task, only underestimate 😅

So to sum up all the tactics of getting better at estimations:

1. Learn from your mistakes
2. Get feedback from other people on your team (cc people on docs aggressively and hold technical review meetings as a forcing function)
3. Don't be afraid to go outside of your team as many projects, especially at Meta, span across multiple teams' services
4. Write stuff down as detailed as you can. Clarity brings accuracy. Try to make docs as high quality as these: System Design Masterclass: Shipping Real Features To Production
5. Really scope out the "extra stuff". This is including (but not limited to): Edge cases, analytics, experimentation details, XFN alignment, security, privacy, legal

Another thing: Yes, it sucks to miss a deadline, but there's layers to it, particularly with proactivity and earliness. If you have a 4 week project and realize after just 1 week that it's a 6 week project, that isn't too bad. However, if you realize in Week 3 that you're going to be late, that's awful. Break all your estimations down into as many milestones as possible, so you can quickly realize when things are off-track.