From Pre-labeling to Production: Engineering Lessons from a Machine Learning Pipeline in the Public Sector DOI: XXXXXXX.XXXXXXX Conference: 5th International Conference on AI Engineering – Software Engineering for AI; April 12–18, 2026; Rio de Janeiro, Brazil ISBN: 978-1-4503-XXXX-X/2018/06
Ronivaldo Ferreira, Guilherme da Silva, Carla Rocha, Gustavo Pinto · 2025
Read From Pre-labeling to Production: Engineering Lessons from a Machine Learning Pipeline in the Public Sector DOI: XXXXXXX.XXXXXXX Conference: 5th International Conference on AI Engineering – Software Engineering for AI; April 12–18, 2026; Rio de Janeiro, Brazil ISBN: 978-1-4503-XXXX-X/2018/06 by Ronivaldo Ferreira, Guilherme da Silva, Carla Rocha, Gustavo Pinto (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-01545.txt.
Description
Ronivaldo Ferreira, Guilherme da Silva, Carla Rocha, and Gustavo Pinto report lessons from running a machine learning pipeline inside Brazil's Brasil Participativo platform, a public participation system. Government settings add burdens private teams rarely face: extreme class imbalance, data drift, bureaucratic access to data, missing dataset versions, and weak control over provenance and monitoring.
Common shortcuts helped and hurt at once. Several shortcuts sped development. Language models did pre-labeling, models were split into routed classifiers, and synthetic data filled gaps. But each shortcut added risks in traceability, reliability, and cost, unless paired with data governance and human checking. The authors argue that responsible machine learning in the public sector is an institutional engineering problem: pipelines are civic infrastructure and must be transparent and reproducible.
Catalog record
| CREATOR | Ronivaldo Ferreira, Guilherme da Silva, Carla Rocha, Gustavo Pinto |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-01545