Gabriel Poesia
Papers
1
Total Citations
14
H-Index
1
About
Gabriel Poesia is a researcher at the forefront of combining large language models with structured algorithmic reasoning. His work centers on enhancing LLMs' capacity for complex, multi-step problem-solving—particularly in program synthesis and formal reasoning. Poesia's most influential contribution is the introduction of **Parsel**, a framework that enables language models to decompose intricate tasks into hierarchical, verifiable subgoals before generating code. This approach directly addresses a critical limitation of LLMs: their struggle with long-horizon planning and logical consistency. By guiding models through a "decompose-then-generate" pipeline, Parsel has demonstrated significant improvements in solving challenging programming problems, achieving a 78% pass rate on APPS introductory-level tasks—a leap from the 60% baseline. This work, already garnering over 14 citations since its 2022 release, has established Poesia as a key voice in neuro-symbolic AI. His research not only advances the practical reliability of LLMs but also offers a principled path toward more interpretable and controllable AI systems, making him a notable figure for students interested in the intersection of natural language processing, automated reasoning, and software engineering.
Research Focus
Key Achievements
Top Papers
- 1