Large Language Models (LLMs) are rapidly reshaping modern software engineering — from intelligent code generation and automated debugging to conversational development environments and AI-assisted programming workflows. However, much of today’s AI-mediated development still relies on informal experimentation, intuition, and ad-hoc prompting practices, often described as “vibe coding.” While effective for rapid prototyping, these workflows frequently lack reproducibility, reliability, security, transparency, and long-term maintainability.

PROMPTOPS 2026 introduces PromptOps as an emerging engineering discipline focused on bringing software engineering rigor to AI-augmented development. The workshop treats prompts, interactions, and developer–LLM workflows as first-class engineering artifacts that can be systematically designed, tested, validated, versioned, governed, and continuously improved.

Unlike venues focused solely on AI coding tools or autonomous agents, PROMPTOPS emphasizes the workflow and process layer of AI-assisted software engineering — bridging human-in-the-loop prompting, execution-grounded validation, tool integration, reproducibility, and trustworthy AI practices.

A key focus of the workshop is the growing importance of multi-turn developer–LLM interaction quality, where conversational structure, contextual evolution, uncertainty, and interaction patterns directly influence software correctness, maintainability, and reliability.

The workshop welcomes research and practical contributions on topics including:

• Prompt engineering and modular prompting strategies
• Prompt testing, validation, and CI/CD pipelines
• Reproducibility and evaluation frameworks
• Developer–LLM conversational workflows
• Prompt traceability and governance
• Hallucination detection and mitigation
• Prompt drift, hidden dependencies, and failure analysis
• Secure and zero-trust AI-assisted development
• Adaptive and self-healing PromptOps workflows

By bringing together researchers, practitioners, tool developers, and industry leaders, PROMPTOPS aims to foster a community dedicated to transforming AI-assisted programming from ad hoc experimentation into a disciplined, reliable, and human-centered software engineering paradigm.

Call for Papers

PROMPTOPS 2026

Engineering Reliable AI-Augmented Software Development

Artificial Intelligence is rapidly transforming modern software engineering. From code generation and debugging to conversational development environments, Large Language Models (LLMs) are reshaping how software is designed, implemented, tested, and maintained. Despite these advances, current AI-assisted development workflows often remain informal, fragile, and difficult to reproduce. Minor variations in prompts or interaction context can lead to inconsistent outputs, raising critical concerns around correctness, reliability, security, maintainability, and governance.

PROMPTOPS 2026 brings together researchers, practitioners, tool developers, and industry leaders to advance the emerging discipline of PromptOps — the engineering principles, workflows, and operational practices required for trustworthy AI-augmented software development.

The workshop aims to move beyond ad hoc “vibe coding” toward rigorous AI-assisted software engineering grounded in:

• Reproducibility and validation
• Human–AI collaboration
• Observability and governance
• Secure and reliable prompting
• Multi-turn developer–LLM workflows
• Execution-grounded software generation and evaluation

PROMPTOPS views prompts, interactions, and conversational workflows as first-class software engineering artifacts that should be systematically designed, tested, versioned, validated, and continuously improved.

---

Call for Papers, Posters & Presentations

We invite high-quality contributions from academia and industry in the form of:

• Research Papers
• Posters
• Presentations

Submissions should advance the science and practice of PromptOps, AI-assisted software engineering, and trustworthy AI-mediated development workflows.

---

Topics of Interest

Topics include, but are not limited to, the following areas:

Prompt Engineering as Software Engineering

• Prompt testing, validation, and CI/CD
• Prompt versioning, modularization, and traceability
• Prompt contracts, schemas, and reusable prompt libraries
• Prompt observability, drift detection, and regression prevention
• Hallucination mitigation and uncertainty-aware prompting
• Prompt governance and lifecycle management

AI-Augmented Development Workflows

• Human–AI collaboration and conversational programming
• Multi-turn developer–LLM interaction patterns
• Execution-grounded prompting and validation loops
• Retrieval-Augmented Generation (RAG) for software engineering
• AI-assisted debugging, repair, and fault localization
• Context management and dialogue-aware software generation
• Multi-modal software engineering workflows

Evaluation, Reproducibility & Benchmarking

• Execution-based evaluation methodologies
• Reproducibility pipelines and artifact evaluation
• Comparative studies across models, prompts, and tools
• Metrics for interaction quality and software reliability
• Failure mode analysis in AI-assisted programming
• Transparent reporting of prompts, datasets, and configurations

Governance, Security & Responsible AI

• Prompt injection defense and adversarial robustness
• Secure and zero-trust prompting
• Compliance-aware PromptOps workflows
• Auditing, logging, and policy enforcement
• Intellectual property and data leakage mitigation
• Responsible, transparent, and trustworthy AI engineering

---

Submission Guidelines

Regular Research Papers

Full papers presenting:

• Novel research contributions
• Empirical studies and evaluations
• Frameworks, methodologies, and tools
• Industrial experiences and deployment studies

Length: Up to 8 pages (excluding references)

---

Posters

Poster submissions may showcase:

• Emerging research ideas
• Prompt engineering platforms and tools
• Reproducibility frameworks
• Developer–LLM interaction studies
• Industrial PromptOps workflows

Submission: 2-page extended abstract (including references and figures)

Accepted posters will be presented during the interactive workshop session.

---

Submission Site

All paper, poster, and presentation proposals must be submitted through the PROMPTOPS 2026 HotCRP submission system.

Submission Portal: https://promptops2026.hotcrp.com

Submissions should be original, unpublished, and not under review elsewhere. Authors are encouraged to carefully follow the workshop submission guidelines before uploading their manuscripts or abstracts.

For any submission-related questions, please contact the workshop organizers.

Call for Presentations

PROMPTOPS 2026 welcomes presentations that encourage discussion, share practical experiences, and showcase emerging advances in PromptOps and AI-assisted software engineering.

We particularly encourage:

Research Presentations

Presentations of published papers, ongoing research, or early-stage ideas related to PromptOps, reproducibility, conversational software engineering, evaluation methodologies, and trustworthy AI-assisted development.

Industrial Experience Reports

Real-world experiences with AI-assisted software engineering, including workflow integration, deployment challenges, security concerns, governance, and lessons learned in practice.

Tool Demonstrations & Tutorials

Demonstrations of PromptOps tools, evaluation frameworks, validation pipelines, prompt CI/CD systems, execution-grounded workflows, and conversational development environments.

Presentation slots are expected to be approximately 20–30 minutes, depending on the number of accepted proposals.

Presentation Proposal Requirements

Presentation proposals should include:

• Title prefixed with “Presentation:”
• Names and affiliations of presenters/authors
• Extended abstract describing the presentation topic and relevance
• Links to related or previously published work (if applicable)
• Brief description of tools/frameworks for demonstrations or tutorials

---

Review Process

All submissions will undergo a rigorous peer-review process conducted by members of the Technical Program Committee and, when appropriate, external reviewers. Each paper submission will receive at least two independent reviews, with additional reviews solicited as needed to resolve discrepancies or ensure adequate expertise.

Submissions will be evaluated based on originality, technical quality, relevance to the workshop themes, empirical rigor, reproducibility, clarity of presentation, and potential impact on AI-augmented software engineering research and practice. Final acceptance decisions will be made by the workshop organizers based on reviewer recommendations and overall program balance.

Poster and presentation proposals will be reviewed for relevance, significance, clarity, and their potential to stimulate discussion within the PROMPTOPS community.

---

Publication & Participation

Accepted papers and posters will be included in the official workshop proceedings, subject to publisher policies.

At least one author of each accepted submission must register and present the work at the workshop.

PROMPTOPS aims to foster a collaborative research community dedicated to building reliable, reproducible, secure, and human-centered AI-augmented software engineering ecosystems.