Does submitting to only journals negatively affect research career after finishing PhD? [D]
Viral discussion questions if skipping top conferences like NeurIPS hurts ML scientist job prospects.
A viral discussion on Reddit's Machine Learning community has sparked a crucial debate about publication strategy for AI researchers seeking industry jobs. The original poster, a PhD graduate or candidate, asked if choosing to publish exclusively in respected journals like TMLR (Transactions on Machine Learning Research), JMLR (Journal of Machine Learning Research), or Neurocomputing—instead of premier conferences like NeurIPS, ICML, or ICLR—would harm their chances of getting interviewed for corporate research scientist positions. This question taps into a long-standing tension in computer science between the fast-paced, prestige-driven conference circuit and the more traditional, peer-reviewed journal model.
Responses from the community, which includes many current industry researchers and hiring managers, revealed a nuanced landscape. A strong consensus emerged that for pure industry research roles at top labs (e.g., Google DeepMind, OpenAI, FAIR), conference publications—especially at NeurIPS, ICML, and ICLR—still carry significant weight. These venues are seen as the primary arenas for cutting-edge work, offering rapid dissemination and high visibility. However, many commenters noted that the substance and impact of the work ultimately matter more than the venue alone. Publishing in a rigorous journal like JMLR is respected, but might be perceived as a slower path.
The discussion highlighted the perceived advantages of journals like TMLR, which was founded to offer a fairer, more transparent review process with opportunities for revision—a contrast to the often random and high-variance reviews at large conferences. For hiring, several users suggested that a strong journal publication could demonstrate deep, polished work, but might need to be supplemented by a visible conference presence or a clear narrative about the research's impact. The safest path for early-career researchers aiming for industry appears to be a mixed portfolio, leveraging conferences for visibility and journals for comprehensive studies.
- Industry hiring for AI scientists still heavily values top conference publications (NeurIPS, ICML, ICLR) for visibility.
- Journals like TMLR and JMLR are respected for rigorous peer review but may be perceived as slower.
- The consensus advises early-career researchers to build a mixed portfolio to maximize both impact and visibility.
Why It Matters
This debate directly shapes the publication strategies and career trajectories of thousands of AI researchers entering the competitive industry job market.