Work in progress: Unlocking code generation through synergistic prompt engineering

Kin-Hon HO; Michael GEORGIADES; Tsz-Kin Justin FAN; Yun HOU; Ken C. K. FONG; Tse Tin David CHAN

doi:10.1109/EDUNINE62377.2025.10980842

Work in progress: Unlocking code generation through synergistic prompt engineering

Kin-Hon HO
, Michael GEORGIADES
, Tsz-Kin Justin FAN
, Yun HOU
, Ken C. K. FONG
, Tse Tin David CHAN

Department of Mathematics and Information Technology (MIT)

Research output: Chapter in Book/Report/Conference proceeding › Chapters

Abstract

Prompt engineering is crucial for optimizing large language models in code generation. This paper explores a synergistic prompt engineering approach that integrates complementary prompting techniques for solving programming problems. Preliminary experiments show that by leveraging the strengths of various prompting techniques, our synergistic approach significantly outperforms traditional single- prompting techniques, improving the accuracy of code generation for Python and C++ exercises. These findings suggest that our synergistic approach is a valuable tool for students, enhancing their interactions with large language models and improving AI-driven programming education. Copyright © 2025 by IEEE.

Original language	English
Title of host publication	Proceedings of IX IEEE World Engineering Education Conference, EDUNINE 2025
Editors	Claudio da Rocha BRITO, Melany M. CIAMPI
Publisher	IEEE
ISBN (Electronic)	9798331542788
DOIs	https://doi.org/10.1109/EDUNINE62377.2025.10980842
Publication status	E-pub ahead of print - 2025

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 4 Quality Education
SDG 9 Industry, Innovation, and Infrastructure

Keywords

Code generation
Prompt engineering

Access to Document

10.1109/EDUNINE62377.2025.10980842

Other files and links

HO, K.-H., GEORGIADES, M., FAN, T.-K. J., HOU, Y., FONG, K. C. K., & CHAN, T. T. D. (2025). Work in progress: Unlocking code generation through synergistic prompt engineering. In C. D. R. BRITO, & M. M. CIAMPI (Eds.), Proceedings of IX IEEE World Engineering Education Conference, EDUNINE 2025 IEEE. Advance online publication. https://doi.org/10.1109/EDUNINE62377.2025.10980842

@inbook{195cf9bb589c49b8a4d34300fcfef4e3,

title = "Work in progress: Unlocking code generation through synergistic prompt engineering",

abstract = "Prompt engineering is crucial for optimizing large language models in code generation. This paper explores a synergistic prompt engineering approach that integrates complementary prompting techniques for solving programming problems. Preliminary experiments show that by leveraging the strengths of various prompting techniques, our synergistic approach significantly outperforms traditional single- prompting techniques, improving the accuracy of code generation for Python and C++ exercises. These findings suggest that our synergistic approach is a valuable tool for students, enhancing their interactions with large language models and improving AI-driven programming education. Copyright {\textcopyright} 2025 by IEEE.",

author = "Kin-Hon HO and Michael GEORGIADES and FAN, \{Tsz-Kin Justin\} and Yun HOU and FONG, \{Ken C. K.\} and CHAN, \{Tse Tin David\}",

year = "2025",

doi = "10.1109/EDUNINE62377.2025.10980842",

language = "English",

editor = "BRITO, \{Claudio da Rocha\} and CIAMPI, \{Melany M.\}",

booktitle = "Proceedings of IX IEEE World Engineering Education Conference, EDUNINE 2025",

publisher = "IEEE",

}

TY - CHAP

T1 - Work in progress: Unlocking code generation through synergistic prompt engineering

AU - HO, Kin-Hon

AU - GEORGIADES, Michael

AU - FAN, Tsz-Kin Justin

AU - HOU, Yun

AU - FONG, Ken C. K.

AU - CHAN, Tse Tin David

PY - 2025

Y1 - 2025

N2 - Prompt engineering is crucial for optimizing large language models in code generation. This paper explores a synergistic prompt engineering approach that integrates complementary prompting techniques for solving programming problems. Preliminary experiments show that by leveraging the strengths of various prompting techniques, our synergistic approach significantly outperforms traditional single- prompting techniques, improving the accuracy of code generation for Python and C++ exercises. These findings suggest that our synergistic approach is a valuable tool for students, enhancing their interactions with large language models and improving AI-driven programming education. Copyright © 2025 by IEEE.

AB - Prompt engineering is crucial for optimizing large language models in code generation. This paper explores a synergistic prompt engineering approach that integrates complementary prompting techniques for solving programming problems. Preliminary experiments show that by leveraging the strengths of various prompting techniques, our synergistic approach significantly outperforms traditional single- prompting techniques, improving the accuracy of code generation for Python and C++ exercises. These findings suggest that our synergistic approach is a valuable tool for students, enhancing their interactions with large language models and improving AI-driven programming education. Copyright © 2025 by IEEE.

UR - https://www.scopus.com/pages/publications/105007412795

UR - https://doi-org.ezproxy.eduhk.hk/10.1109/EDUNINE62377.2025.10980842

U2 - 10.1109/EDUNINE62377.2025.10980842

DO - 10.1109/EDUNINE62377.2025.10980842

M3 - Chapters

AN - SCOPUS:105007412795

BT - Proceedings of IX IEEE World Engineering Education Conference, EDUNINE 2025

A2 - BRITO, Claudio da Rocha

A2 - CIAMPI, Melany M.

PB - IEEE

ER -

Work in progress: Unlocking code generation through synergistic prompt engineering

Abstract

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint