Building Your Personal Brand and Disseminating Individual Works
In today’s academic and technical fields, conducting high-quality research is only part of the journey. Equally important is disseminating your work effectively—making sure your contributions are seen, understood, and built upon by others. This not only increases the impact of your work but also strengthens your profile, opens doors for collaborations, and fuels your career growth.
This post focuses on research dissemination within the academic community especially in the field of Artificial Intelligence (AI) / Machine Learning (ML). The insights are based on my personal experience, observations, and input from Yarin Gal. It is intended for PhD students and early-career researchers such as postdocs.
This post approaches research dissemination from two perspectives: building your personal brand and promoting individual works. These two aspects complement each other—a strong personal brand facilitates the dissemination of your work, while sharing your work actively reinforces your brand.
The first step in building a strong and discoverable academic presence is to create public profiles on major platforms and keep them consistently updated and in sync. At a minimum, you should have active profiles on Personal website, X (Twitter), LinkedIn, Google Scholar, GitHub, ORCID.
Each serves a unique purpose, but your personal website should be the central hub that links out to all other profiles—and vice versa. It allows others to easily find all your professional information—such as your research, code, talks, and contact details—in one place. Setting up a personal website is easier than ever. GitHub Pages offers a free and customizable way to host your site. Some popular templates to get started quickly include Al-folio and AcademicPages.
Next, focus on improving your visibility within the academic community. The goal is for others to recognize your name and associate it with your expertise—even without a self-introduction—and to think of you when seeking advice and/or work in your area of specialisation.
Broadly speaking, there are four communities you should care about: your subfield(s), your broader field, the academic community at large, and the general public. Start by identifying which communities you want to target, as each requires different strategies. This post focuses on the first two: your subfield and the broader field. To illustrate the distinction, think of your subfield community as the group of people who attend the same workshops, while your broader field includes all researchers present at major conferences like ICLR/ICML/NeurIPS in AI.
The foundation of your reputation in the community is your research. High-quality work speaks for itself and for you—it advances your field and draws the attention of your peers. While not everyone can produce groundbreaking results, a more attainable path is to delve deeply into a specific topic and consistently publish solid, well-regarded work. This strengthens your chances of being recognised as an expert in that area. Continuity is key—it demonstrates that you are an active contributor, not a figure of past relevance.
Collaboration offers a unique opportunity for your peers to get to know not only your research, but also you—your communication style, reliability, and work ethic. In fact, this personal impression can be even more influential than your published work. Collaborators often serve as informal ambassadors of your reputation; they may be asked for their opinions about you, or naturally share their experiences with others in the field. These impressions can travel far and shape how you’re perceived in both academic and professional circles. That’s why it’s important to actively seek collaborations and approach each one with professionalism and commitment. A single negative experience can undermine your credibility, while a strong collaboration can significantly elevate your standing and open doors to new opportunities.
Engagement and Networking. Engagement involves participating in community events, while networking refers to building relationships with fellow members of the field. The core aim is to establish professional or social connections. These interactions can take various forms—virtual (e.g. social media, online forums, virtual seminars) or in-person (e.g. conferences, workshops, lab visits), professional (e.g. academic panels and talks) or social (e.g. informal meetups, conference dinners). By actively engaging in these settings, you increase your visibility, stay up to date with new developments, and create opportunities for collaboration. These interactions not only help others discover your work but also give you the right moments to introduce it—whether through invited talks, seminar presentations, or informal conversations. Over time, consistent and authentic engagement helps establish you as a recognised and trusted member of the community. In fact, when you see this blog, we are interacting! If you find this useful hopefully my visibility will be raised in your mind.
Contributing to academic services—such as reviewing papers, serving on program committees, or organizing workshops—is a valuable way to build your research brand. These roles position you as an active and trusted member of the community, deepen your understanding of the field, and increase your visibility among peers and senior researchers. Reviewing papers sharpens your critical thinking and often leads to recognition from conference organizers or editors, which can open doors to further opportunities like serving as an area chair or editor. Organizing events or co-chairing sessions allows you to shape the conversation in your research area and demonstrate leadership. Over time, consistent service builds a reputation for reliability, expertise, and commitment, which enhances both your credibility and your influence in the academic ecosystem.
This section discusses some practical tips for disseminating your work throughout its lifecycle. The key idea behind all of them is simple: understand how people in the target community you want to reach discover new research, and make sure your work shows up attractively and accessibly in those channels.
The first step in disseminating your research is to make it easy to disseminate. So, what kind of work is easy to disseminate? In general, it should be Attractive, Reliable, and Accessible.
A research work is attractive if, even when summarised in just 2-3 sentences, it makes people want to learn more. Ideally, it tackles an important or interesting problem in the field and delivers useful or surprising findings. Of course, there’s no universal definition of what counts as “important,” “interesting,” or “useful”—so you’ll need to rely on your own research taste and try to anticipate how the community would react to the work, independent of who wrote it or where it’s published. Neel Nanda has a great post on cultivating research taste, and Chris Olah offers some helpful exercises to sharpen it.
There’s also a bit of a “cheat code”: do timely research. That means working on a hot topic where community interest is obvious and the space is still relatively unexplored. Think LLMs in 2023, LLM-based agents in 2024, or reasoning models in early 2025. The key here then becomes speed—if you move fast, you can explore new questions before the space gets saturated.
Once you’ve captured people’s attention, they’ll likely start digging into the details—this is where reliability matters most. You need to make sure your work stands up to scrutiny: the theory should be solid, the experiments well-designed, the results reproducible, and the conclusions fair and grounded. After all, sharing something that’s incorrect or overhyped isn’t just risky—it feels like publicly embarrassing yourself.
Now, assuming people are interested in using your work, the final push is to make it accessible. This means open-sourcing both your code and model weights, so others can easily build on what you’ve done. Try to structure your code well—make it modular and turn it into tools that are easy to plug into other projects. To further boost accessibility, consider integrating your code into commonly used libraries or toolboxes in your field. For example, in adversarial ML, researchers often use frameworks like Torchattacks or ART to run evaluations across various attack methods. Getting your method included in these toolboxes makes it much easier for others to adopt your work with minimal effort.
Of course, you should aim to publish your work at the most prestigious venue you can—these naturally bring more visibility, as people in the field are always looking out for the latest progress from top conferences and journals. Compared to journals, workshops and conferences also offer the chance to present your work in person and connect directly with peers.
In AI/ML, a high-exposure publication path often looks like this: arXiv → workshop → conference → journal. The workshop should be non-archival, so it doesn’t block you from submitting to a formal venue later. If you go on to publish in a journal, the work has to be a substantial extension of your earlier conference version.
Beyond conventional publication, you can also boost visibility by submitting your work to public challenges, benchmarks, or model zoos. These platforms often attract a wide audience and can help your work gain traction in both academic and applied communities. Another option is to “submit” your work to ongoing surveys on arXiv like this or curated paper collections on GitHub, like this one. I’ve personally received self-recommendations like this for our own survey paper, and my collaborators and I are usually happy to add something if it meets our criteria. That said, use this approach thoughtfully and politely; if done carelessly, it can come across as annoying.