How to Network at Academic Conferences: A Graduate Student's Guide

1. Prepare Before You Arrive

Effective conference networking starts two weeks before the event. Read through

the program and identify 5-10 researchers whose work directly relates to yours.

Look them up — read their two most recent papers, check their current projects,

and note any open questions their work raises. When you approach someone at a

conference and say "I read your paper on X and had a question about Y," you

immediately stand out from the dozens of people who approach them with "Hi, I'm

a PhD student, nice to meet you." Also email 2-3 people in advance to set up

brief coffee meetings — many senior researchers appreciate the organized approach

and will respond.

2. Master the 90-Second Introduction

When someone asks "what do you work on?", you have about 90 seconds before

their attention drifts. Prepare a layered answer: one sentence for the

non-expert ("I study how machine learning models fail in medical imaging"),

three sentences for a specialist in your field, and five minutes for a

deep conversation. Do not lead with your institution or advisor's name —

lead with the problem you are solving. The goal is to say something interesting

enough that the other person asks a follow-up question. Practice this

introduction out loud before the conference; it sounds different than it reads.

3. Work the Poster Session Strategically

Poster sessions are the best networking opportunity at any conference —

better than talks, because conversations are two-way. Spend the first 20

minutes walking the entire session to identify which posters you want to

visit. Return when the presenter is not surrounded by three rows of people.

When presenting your own poster, avoid reading from it. Instead, give a

60-second oral summary and then ask the visitor what they are working on.

The best poster conversations end with both people learning something new.

Keep a small notebook or use your phone's notes app to record names and

what you discussed — memory decays fast across a full conference day.

4. Navigate Meals and Social Events

Conference dinners and social events are networking opportunities, but they

require a different approach. Do not sit with your labmates at dinner — you

already know them. Introduce yourself to whoever is next to you in the food

line and sit with people you have not yet met. At the welcome reception, aim

to have 4-5 short conversations (10-15 minutes each) rather than one long

conversation with the same person. Move on gracefully with "I do not want to

monopolize your time — I hope we get a chance to talk more later." This

leaves both parties with a positive impression and keeps the conversation open.

5. Follow Up Within 48 Hours

The connections you make at a conference are fragile — they fade quickly

without reinforcement. Within 48 hours, send a brief personalized email to

each person you had a meaningful conversation with. Reference something

specific: "It was great to hear about your work on sparse attention mechanisms.

I looked up the paper you mentioned — the Figure 3 results were really

surprising to me." Attach your paper or a relevant preprint if appropriate.

Connect on LinkedIn or follow on Twitter/X only if you plan to actually engage

with their posts. An empty connection is worse than no connection.

6. Build Long-Term Academic Relationships

Networking is not transactional — it is about building relationships that

develop over years, not days. After the conference, engage with people's

work: cite their papers when appropriate, comment thoughtfully on their

preprints, and share their work when it is relevant to your network.

When you publish something new, send a personal email to the 3-5 researchers

it is most relevant to. Over time, these interactions create the visibility

and mutual respect that lead to collaboration invitations, recommendation

letters, and job opportunities. The graduate students who build strong

networks are not necessarily the most extroverted — they are the most

consistent.