There’s something very surreal about stepping into your first major machine learning conference: suddenly, all those GitHub usernames, paper authors, and protagonists of heated twitter spats become real people, the hallways are buzzing with discussions of papers you’ve been meaning to read, and somehow there are 17,000 other people trying to navigate it all alongside you. That was my experience at NeurIPS this year, and despite feeling like a microplankton in an ocean of ML research, I had a grand time. While some of this success was pure luck, much of it came down to excellent advice from the group’s ML conference veterans and lessons learned through trial and error. So, before the details fade into a blur of posters and coffee breaks, here’s my guide to making the most of your first major ML conference.

Navigating the Science

• Prioritise workshops. The workshops, at least for those more application-ly inclined, are absolutely where it’s at, so don’t feel rubbish if you reach the end of the main conference track feeling like you haven’t gained much. For NeurIPS specifically, highlight workshops were Machine Learning for Structural Biology and AIDrugX – both of these report quite high acceptance rates, so if you’re set on attending one of these conferences but find applying to the main track a bit daunting, I’d strongly recommend going for one of them.
• Plan your poster sessions. If you’re used to normal-scale conferences, you might think you can find posters of interest by strolling about and window shopping. Spoiler alert: you cannot do this at a conference that 17,000 people are attending! I gained very little from the poster sessions I hadn’t prepared for. While there’s some clustering, it’s not a perfect system, and I certainly missed interesting work I would have caught if I’d gone through the paper list beforehand.
• Navigate the grouping system. One of the tricky parts of these mega-conferences is that the research spans both core ML development and its applications, which leads to some interesting categorisation challenges. For instance, I found relevant papers scattered across different sections – some were grouped by field (like bioinformatics), while others were organised by methodology (eg in the diffusion models or computer vision sections).
• Don’t skip the expo hall. The larger booths within the expo halls had their own talk schedules that weren’t a part of the main conference listing – I’d recommend stopping by as soon as you arrive so you can plan to include these in your scheduling efforts.
• Originality is all you need. There were 11 papers accepted to the conference’s main track ending in ‘is all you need’. And that’s all I’ll say on that.

Jobs and Networking

• Update your CV before you go. Trust me, you’ll thank yourself later when you’re having a conversation with a potential employer and can confidently (or at least not ashamedly) share your most up-to-date version.
• Connect immediately after meeting people. Add/follow people to your LinkedIn/Twitter/X/Bluesky/whatever your poison as soon as you’ve met them. Don’t wait until you’re back home – trying to find ‘that guy with the monosyllabic name wearing glasses and a fleece who was interested in generative models’ at an ML conference is like trying to find a very specific needle in a stack of very similar needles.
• Always register your interest. Don’t skip sign-ups to register your interest with companies you might want to work for. At these big conferences, a lot of the best networking happens at private company events that you’ll only hear about if you’ve registered your interest on a form or chatted with someone (also, many of these registration forms will ask for your CV – remember that updated version from earlier?)

Practical Tips for Survival

• Find your conference buddies. Despite being an advocate for solo adventures (if you saw me watching Paddington 3 at the cinema alone – no, you didn’t), I would have found this conference challenging without a few reliable coffee and meal companions. Not everyone will be as fortunate as I was to attend with the most excellent OPIGlets Nele and Nikhil, but having conference buddies is invaluable. Beyond the social aspect, it’s particularly helpful for the evening events – having someone to walk home with in a foreign city late at night is nice (/essential, depending on your personal demographic). Don’t worry if you don’t know anyone before you set off – the scale of these conferences means it is almost impossible to avoid meeting at least a few other attendees either on your travels to the conference or in your accommodation.
• Know when to take breaks. This might be more of a personal take, but if you have a slot of the conference where none of the talks are particularly relevant to you or you feel too tired to focus, call it, and go and have a break – there’ll always be something on later and I found doing fewer things more well rested to be better than spreading myself too thin. The conference is a marathon, not a sprint, and you’ll get more value from being fully present for the sessions you do attend than zombie-walking through everything.
• Master the app. The conference app is your best friend. Download it, learn it, love it. It’s essential for keeping track of your schedule, finding rooms, and staying updated on any last-minute changes.
• Be selective with swag. When visiting the expo hall, remember that everything you pick up needs to fit in your return luggage. Maybe this comes more naturally to those with weaker hunter-gatherer instincts than I, but trying to layer two hats from unnamed quant research companies on departure day because they wouldn’t fit in my checked bag looked exactly as ridiculous as it sounds.

So if you’re gearing up for your first big ML conference, remember: you don’t have to do everything perfectly. Just come prepared, be strategic about your time, and don’t forget to enjoy the controlled chaos of it all. Everyone there was a first-timer once – even if some of them pretend they emerged from the womb clutching a NeurIPS badge and a copy of the original transformer paper.