Peter C. Rigby

Yuecai Zhu, Nikolaos Tsantalis, Peter C. Rigby

The promise of Large Language Models in automated software engineering is often measured by functional correctness, overlooking the critical issue of long term maintainability. This paper presents a systematic audit of technical debt in AI-generated software, revealing that AI does not eliminate flaws but rather introduces a distinct machine signature of defects. Our multi-scale analysis, spanning single-file algorithmic tasks and complex, agent generated systems, identifies a fundamental Reasoning-Complexity Trade-off: as models become more capable, they generate increasingly bloated and coupled code. This architectural decay is so pronounced that we establish a Volume-Quality Inverse Law, where code volume is a near perfect predictor of structural degradation. Crucially, we demonstrate that neither functional correctness nor detailed prompting mitigates this decay. These findings challenge the current paradigm of prompt-driven generation, reframing the central problem of AI-based software engineering from one of code generation to one of architectural complexity management. We conclude that future progress depends on equipping agents with explicit architectural foresight to ensure the software they build is not just functional, but also maintainable.

Chandra Maddila, Adam Tait, Claire Chang et al.

Aim: With the advent of LLMs, sophisticated agentic program repair has become viable at large organizations with large codebases. In this work, we develop an Engineering Agent that fixes the source code based on test failures at scale across diverse software offerings internally. Method: Using Llama as the base, we employ the ReAct harness to develop an agent. We start with a test failure that was triaged by a rule-based test failure bot. We then set up an agentic harness and allow the agent to reason and run a set of 15 actions from reading a file to generating a patch. We provide feedback to the agent through static analysis and test failures so it can refine its solution. We leverage an LLM-as-a-Judge to ensure that the patch conforms to the standards followed by a human review to land fixes. Benchmark Findings: We curated offline benchmarks for our patch generator, the Engineering Agent loop, and the LLM-as-a-Judge. In offline evaluations we found that a specialized 70B model is highly competitive with the much larger but vanilla Llama-405B. In an ablation study, we found that the ReAct harness (neural model) benefited from the symbolic information from static analysis tools and test execution traces. A model that strikes a balance between the solve rate and error rate vs the cost and latency has a benchmark solve rate of 42.3% using an average 11.8 feedback iterations. Production Findings: In a three month period, 80% of the generated fixes were reviewed, of which 31.5% were landed (25.5% of the total number of generated fixes). Feedback from Engineers: We used open coding to extract qualitative themes from engineers' feedback. We saw positive feedback in the form of quick approvals, gratitude, and surprise. We also found mixed feedback when the Engineering Agent's solution was partially correct and it served as a good starting point.

Ali Sayedsalehi, Peter C. Rigby, Gregory Mierzwinski

Performance regression testing is essential in large-scale continuous-integration (CI) systems, yet executing full performance suites for every commit is prohibitively expensive. Prior work on performance regression prediction and batch testing has shown independent benefits, but each faces practical limitations: predictive models are rarely integrated into CI decision-making, and conventional batching strategies ignore commit-level heterogeneity. We unify these strands by introducing a risk-aware framework that integrates machine-learned commit risk with adaptive batching. Using Mozilla Firefox as a case study, we construct a production-derived dataset of human-confirmed regressions aligned chronologically with Autoland, and fine-tune ModernBERT, CodeBERT, and LLaMA-3.1 variants to estimate commit-level performance regression risk, achieving up to 0.694 ROC-AUC with CodeBERT. The risk scores drive a family of risk-aware batching strategies, including Risk-Aged Priority Batching and Risk-Adaptive Stream Batching, evaluated through realistic CI simulations. Across thousands of historical Firefox commits, our best overall configuration, Risk-Aged Priority Batching with linear aggregation (RAPB-la), yields a Pareto improvement over Mozilla's production-inspired baseline. RAPB-la reduces total test executions by 32.4%, decreases mean feedback time by 3.8%, maintains mean time-to-culprit at approximately the baseline level, reduces maximum time-to-culprit by 26.2%, and corresponds to an estimated annual infrastructure cost savings of approximately $491K under our cost model. These results demonstrate that risk-aware batch testing can reduce CI resource consumption while improving diagnostic timeliness. To support reproducibility and future research, we release a complete replication package containing all datasets, fine-tuning pipelines, and implementations of our batching algorithms.

Chandra Maddila, Negar Ghorbani, James Saindon et al.

Aim. There are 10s of thousands of code review comments each week at Meta. We developed Metamate for Code Review (MetaMateCR) that provides AI-assisted fixes for reviewer comments in production at scale. Method. We developed an internal benchmark of 64k <review comment, patch> data points to fine-tune Llama models. Once our models achieve reasonable offline results, we roll them into production. To ensure that our AI-assisted fixes do not negatively impact the time it takes to do code reviews, we conduct randomized controlled safety trials as well as full production experiments. Offline Results. As a baseline, we compare GPT-4o to our small and large Llama models. In offline results, our LargeLSFT model creates an exact match patch 68% of the time outperforming GPT-4o by 9 percentage points (pp). The internal models also use more modern Hack functions when compared to the PHP functions suggested by GPT-4o. Safety Trial. When we roll MetaMateCR into production in a safety trial that compares no AI patches with AI patch suggestions, we see a large regression with reviewers taking over 5% longer to conduct reviews. After investigation, we modify the UX to only show authors the AI patches, and see no regressions in the time for reviews. Production. When we roll LargeLSFT into production, we see an ActionableToApplied rate of 19.7%, which is a 9.2pp improvement over GPT-4o. Our results illustrate the importance of safety trials in ensuring that AI does not inadvertently slow down engineers, and a successful review comment to AI patch product running at scale.

Vijayaraghavan Murali, Chandra Maddila, Imad Ahmad et al.

Generative LLMs have been shown to effectively power AI-based code authoring tools that can suggest entire statements or blocks of code during code authoring. In this paper we present CodeCompose, an AI-assisted code authoring tool developed and deployed at Meta internally. CodeCompose is based on the InCoder LLM that merges generative capabilities with bi-directionality. We have scaled up CodeCompose to serve tens of thousands of developers at Meta, across 9 programming languages and several coding surfaces. We present our experience in making design decisions about the model and system architecture for CodeCompose that addresses these challenges. To release a LLM model at this scale, we needed to first ensure that it is sufficiently accurate. In a random sample of 20K source code files, depending on the language, we are able to reproduce hidden lines between 40% and 58% of the time, an improvement of 1.4x and 4.1x over a model trained only on public data. We gradually rolled CodeCompose out to developers. At the time of this writing, 16K developers have used it with 8% of their code coming directly from CodeCompose. To triangulate our numerical findings, we conduct a thematic analysis on the feedback from 70 developers. We find that 91.5% of the feedback is positive, with the most common themes being discovering APIs, dealing with boilerplate code, and accelerating coding. Meta continues to integrate this feedback into CodeCompose.

Andrey Krutauz, Tapajit Dey, Peter C. Rigby et al.

Aim: In contrast to studies of defects found during code review, we aim to clarify whether code reviews measures can explain the prevalence of post-release defects. Method: We replicate a study by McIntoshet. al that uses additive regression to model the relationship between defects and code reviews. To increase external validity, we apply the same methodology on a new software project. We discuss our findings with the first author of the original study, McIntosh. We then investigate how to reduce the impact of correlated predictors in the variable selection process and how to increase understanding of the inter-relationships among the predictors by employing Bayesian Network (BN) models. Context: As in the original study, we use the same measures authors obtained for Qt project in the original study. We mine data from version control and issue tracker of Google Chrome and operationalize measures that are close analogs to the large collection of code, process, and code review measures used in the replicated the study. Results: Both the data from the original study and the Chrome data showed high instability of the influence of code review measures on defects with the results being highly sensitive to variable selection procedure. Models without code review predictors had as good or better fit than those with review predictors. Replication, however, confirms with the bulk of prior work showing that prior defects, module size, and authorship have the strongest relationship to post-release defects. The application of BN models helped explain the observed instability by demonstrating that the review-related predictors do not affect post-release defects directly and showed indirect effects. For example, changes that have no review discussion tend to be associated with files that have had many prior defects which in turn increase the number of post-release defects.