About

Getting to the talks that were always just out of reach

I have been to AWS re:Invent several times, and every one of them was worth the trip. But every one of them also ended with a shortlist of sessions I had marked, wanted, and never got to — not because they clashed, but because nothing could tell me in time that they were within reach. I wanted those hours back, and I wanted them spent on the good stuff.

The week is already good. The question that kept nagging me was how much better it could be if the logistics were solved rather than improvised — if the chalk talk one building over showed up as an option instead of staying invisible until the day after. So I built a planner for myself, in evenings and weekends, and then handed it to the colleagues I travel with, because they had the same shortlist problem I did.

Three things that kept good sessions out of reach

Two sessions that fit the calendar and not the map

The one you only find out about on the day. Two talks you want, back to back, and the calendar says they fit. They do not fit: the venues are kilometres apart, the walk inside each resort is longer than the walk between them, and sixty thousand other people are in the corridor with you. You arrive ten minutes into a room that filled at the door, and an hour that was going to be useful is now a walk. Nothing warns you in advance, because as far as a calendar is concerned two non-overlapping blocks are simply fine — a calendar has no idea there is a mile between them. And the version that would have worked, the same talk repeated on Thursday at a venue you are already at, is sitting in the catalogue the whole time with nothing pointing you at it.

A catalogue too big to read and too flat to search

Thousands of sessions, seven filters and a search box. The talks that would have been the highlight of my week were in there — the search simply had no way to hand them to me, because reaching a thing requires already knowing the word for it. Prepare as thoroughly as you like and the tool still rewards you for recognising titles rather than for surfacing what you did not know to look for. That is a lot of good content nobody ever finds.

The gaps that cannot be filled in four minutes

A free afternoon looks like breathing room in November. In December it is a puzzle with a four-minute deadline: what is on, and reachable from where you are standing right now, and worth the hour. Nobody solves that on a phone in a corridor between two rooms, so the hour goes to whatever is nearest and familiar. Which is fine — except there was very likely a chalk talk one building over that would have been the best hour of the day, and it was reachable, and nothing said so.

None of that is a catalogue problem

The official session catalogue is a good catalogue. It lists what is on, it lets you reserve a seat, and it is accurate. It just is not a planner, and it was never trying to be. A catalogue answers “what exists?”. What I needed answered was “given where I am, what I care about, and what I have already booked — what should I do next, and can I actually get there?”

That question needs three things a catalogue does not have: a model of the walk between rooms, a model of what you personally care about, and your own schedule to measure both against. So that is what I built.

So I made them reachable

• The walk is a hard constraint, not a footnote. The planner walks a route graph between rooms rather than venues. Every result carries a feasibility badge measured against the block before it in your own day, every booked block carries the walk in front of it, and reminders fire when it is time to leave rather than when the talk starts.
• You ask in a sentence. Type what you are after and the AI reads it into the same filters the catalogue has — shown back as chips you can correct or drop, so you can see what it did with your words instead of trusting it.
• A gap is not left blank. An empty afternoon lists the venues genuinely reachable from where your last session leaves you, and suggests what is worth your time there, with a reason on every row.

Every feature in the app is something I wanted on a specific afternoon in Las Vegas, standing somewhere, holding a phone. There is no roadmap of features nobody asked for.

The point of all of it is not to save time for its own sake. It is that the same five days hold noticeably more of what you came for: the second talk you can now make because the repeat was offered, the chalk talk that turned up in a gap you would have spent walking back to somewhere familiar, the session you would never have searched for because you did not have the word for it yet. Same trip, more of the good stuff in it.

Who is behind it

I am Jim van Eijk, a Dutch 🇳🇱 Frontend developer working at a software company 23G. I’ve been to re:Invent for work. This planner started as a way to get more out of the trip we were already taking, and it is still what my colleagues and I use out there. But it is a personal project, built outside 23G's hours. 23G is the reason I am there; it is not the company shipping this.

Which also means there is no support desk. There is me, and an email address, and a genuine interest in hearing which part of your week it failed to help with.

What this is not

• Not official. SessionStack is an independent, unofficial planner, not affiliated with or endorsed by Amazon Web Services. AWS and re:Invent are trademarks of Amazon.com, Inc. or its affiliates, used here only to say which event this plans.
• Not a replacement for registering. You still register, and you still reserve your seats, through AWS. This plans the week around what you have.
• Not watching you. No analytics, no advertising, no tracking SDKs and no third-party profiling in the app, and nothing sold or shared for marketing. The privacy policy spells out what is stored and why.

Tell me what is missing

If you are going to re:Invent and the app does not do the thing you need it to do, I would rather hear it than not: jim@23g.nl. The iOS build is still in testing: join the TestFlight beta .

If you are going for the first time, the first-timer guide is most of what I wish somebody had told me, and the session formats matter more than the titles do.