We are excited to announce the Industry Track at COLING 2027, which will be held in Macau, China on May 9 – 14, 2027. Language technologies and their practical applications, particularly with the recent popularity and adoption of Large Language Models (LLMs), have deeply permeated many aspects of our daily lives. Though the origins of these advancements can often be traced back to academic and industrial research facilities, translating these innovations into industrially robust applications presents an entirely distinct set of technical hurdles. We are soliciting contributions that explore novel ideas and implementations across all domains of speech and natural language processing technologies, particularly LLMs, and systems pertinent to practical scenarios. This track prioritizes submissions that deepen our comprehension and showcase effective strategies for managing the practical difficulties associated with deploying language processing technologies in complex real-world environments. By “complex real-world system,” we refer to applications that are operational outside of controlled environments and incorporate natural language processing and the potential use of LLMs. The emphasis in this track is not on research novelty, but on the practical application of the technologies in real-world settings taking into account computational limitations, ethical considerations, novel applications, model freshness, etc. It is not mandatory that these systems be produced by commercial enterprises; however, their user base should extend beyond the confines of the NLP research community. *Important Dates* - Paper submission deadline: November 30, 2026 - Notification of acceptance: March 15, 2027 - Main conference, including industry track: May 9 – 14, 2027 All deadlines are 11.59 pm UTC -12h (anywhere on earth). *Topics of Interest* The COLING 2027 Industry Track highlights key insights, operational practices, and emerging research challenges arising from the practical deployment of speech and natural language processing (NLP) systems, including Large Language Models (LLMs). We invite submissions across all areas of speech and natural language processing, grouped into three main core themes: *1. System Design, Infrastructure & Operational Efficiency* - *Efficiency & Latency*: Optimization techniques for real-world constraints, including low-latency inference, model compression, quantification, and cost reduction for LLM/LM training and deployment at scale. - *Continuous Lifecycle & Maintenance*: Longitudinal studies of systems in production, tracking performance drift, maintenance burdens, model update strategies, and model freshness over time. - *Data & Model Governance*: Procedures for deployed systems, including versioning, data lineage, compliance, access control, auditing, and corporate regulatory constraints. - *Enabling Infrastructure*: Frameworks, orchestration, and system design choices (e.g., offline vs. online, hybrid symbolic/neural setups, edge computing) for large-scale production deployment. *2. Deployed Methods & System Evaluation* - *Evaluation & Quality Assurance*: Methodologies for testing LLM-based components in production, in situ evaluation, automated vs. human-in-the-loop evaluation, and handling non-deterministic system behavior. - *Robustness & Trustworthiness*: Safety, harmlessness, bias mitigation, fairness, interpretability, and handling unexpected or adversarial user behavior in production environments. - *Human-in-the-Loop & Interactive Systems*: Practical designs for human-AI interaction, active learning, human feedback integration (e.g., RLHF/DPO in production), and user satisfaction tracking. - *Negative Results & Trade-Offs*: Lessons learned from failed deployments, architectural trade-offs, and critical retrospective analysis of real-world system designs. *3. Novel Applications, Case Studies & Resources* - *Real-World Case Studies*: End-to-end case studies from problem formulation to deployment, highlighting business impact, market opportunity, trade-offs, and actionable insights. - *Domain-Specific Applications*: Novel real-world applications across domains such as healthcare, finance, legal, e-commerce, software engineering, and public services/citizen-facing systems. - *Application-Relevant Datasets*: Design, curation, annotation, and sanitization of specialized or domain-specific real-world datasets. - *Best Practices & Vision Pieces*: Opinion papers, vision pieces, and actionable summaries of best practices in practical language technology adoption. *Submission* Authors are invited to submit original, full-length (6 page) industry track papers that are not previously published, accepted to be published, or under consideration for publication in any other forum. Manuscripts should be submitted electronically, in PDF format and formatted using COLING 2027 formatting requirements. Please do not modify these style files, and do not use templates designed for other conferences. Submissions that do not conform to the required styles, including paper size, margin width, and font size restrictions, will be rejected without review. Industry Track papers cannot exceed 6 pages in length; however, references do not count toward the 6-page limit, nor do the following optional sections: acknowledgements (only in the final version), ethical considerations, and appendices. Appendices may include, for example, pre-processing decisions, model parameters, feature templates, pseudocode, sample system inputs/outputs, information about user studies, additional errors analysis and other details that are necessary for the replication of the work described in the paper. In addition, authors may submit a separate file with supplementary materials such as software or data. Note that the 6-page paper needs to remain fully self-contained: appendices and supplementary material are completely optional, and reviewers are not even asked to review or download them. Papers are submitted via the COLING 2027 Industry Track system <https://openreview.net/group?id=coling-iccl.org/COLING/2027/Industry_Track> . *Double-blind Review* Starting with COLING 2027, the Industry Track has transitioned from single-blind to double-blind reviewing to align with ACL community standards and ensure an equitable review process. To facilitate this, submissions must neither include the authors’ names nor their affiliations. ============ *Full CfP*: https://2027.coling-iccl.org/calls/industry_track/ *Reviewers and ACs expression of interest*: https://forms.gle/BDsNxuddSCRqrbnE7 *Contact*: coling2027industry@gmail.com