Teaching Software Engineering with LLM and MCP Integration: From Classroom to Industry Practice

Jacky Keung; Kehui Chen; Weining Li; Xiangbing Shao; Xiaoxue Ma; Yishu Li

arxiv: 2606.19167 · v1 · pith:SCX2754Bnew · submitted 2026-06-17 · 💻 cs.SE

Teaching Software Engineering with LLM and MCP Integration: From Classroom to Industry Practice

Kehui Chen , Jacky Keung , Weining Li , Xiangbing Shao , Yishu Li , Xiaoxue Ma This is my paper

Pith reviewed 2026-06-26 20:19 UTC · model grok-4.3

classification 💻 cs.SE

keywords software engineering educationlarge language modelsmodel context protocolcollaborative teaching modelindustry internshipsAI-assisted learningeducational innovation

0 comments

The pith

Integrating LLMs and the Model Context Protocol into software engineering teaching bridges classroom instruction with industrial workflows.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper presents a teaching model that integrates large language models and the Model Context Protocol into software engineering education. The model embeds these tools into daily teaching, code assistance, and engineering simulations to connect classroom learning with industrial workflows. It also incorporates industry internships to allow students to apply these technologies in real settings. If effective, this would update education to match current industry demands involving AI tools.

Core claim

The research investigates an innovative approach that integrates LLMs and MCP into a collaborative teaching model for software engineering education. By embedding LLM and MCP driven tools into daily teaching, code assistance, and engineering simulations, the model bridges the gap between traditional instruction and industrial workflows. This enhances students' programming competence, practical problem-solving abilities, and proficiency in using intelligent engineering tools. Through industry internships, students apply these technologies in real-world settings, strengthening the connection between academic preparation and professional practice.

What carries the argument

The collaborative teaching model that embeds LLM and MCP driven tools into daily teaching, code assistance, and engineering simulations, supported by industry internships.

If this is right

Students gain enhanced programming competence through daily use of the tools.
Practical problem-solving abilities strengthen via engineering simulations.
Proficiency in intelligent engineering tools increases with classroom integration.
The link between academic preparation and professional practice becomes stronger through internships.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The model could be tested for scalability across different university sizes or regions.
Instructor training requirements for effective LLM and MCP use might need explicit development.
Similar tool integrations could be explored in adjacent fields like data engineering education.

Load-bearing premise

That embedding LLM and MCP tools along with industry internships will produce measurable gains in student competence and bridge the academic-industry gap without specified controls or outcome metrics.

What would settle it

A controlled comparison of specific student outcomes such as project completion rates, problem-solving test scores, or tool proficiency assessments between classes using the integrated model and traditional instruction.

read the original abstract

The rapid integration of Large Language Models (LLMs) and the Model Context Protocol (MCP) into industrial software engineering has created a pressing need to update software engineering education to align with emerging technologies and evolving industry demands. This study investigates an innovative approach that integrates LLMs and MCP into a collaborative teaching model for software engineering education, aiming to build a practical learning framework closely connected to real-world engineering practices. By embedding LLM and MCP driven tools into daily teaching, code assistance, and engineering simulations, the model effectively bridges the gap between traditional instruction and industrial workflows. This integration enhances students' programming competence, practical problem-solving abilities, and proficiency in using intelligent engineering tools. Furthermore, through partnerships with industry internships, students can apply these technologies in real-world settings, further strengthening the connection between academic preparation and professional practice. Overall, this research offers a practical pathway for reforming and innovating software engineering education in the era of artificial intelligence.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This is a curriculum proposal for folding LLMs and MCP into software engineering classes plus internships, but it asserts skill gains without any data or evaluation to back them.

read the letter

This paper describes a model for teaching software engineering that incorporates LLMs and the Model Context Protocol both in class and through industry internships. The key point is that it claims this setup bridges academic and industrial practice and improves student skills, yet it provides no data or evaluation to support those claims.

The new element here is the concrete framing around MCP integration in daily teaching activities and the direct link to internships that use the same tools. It gives a clear picture of how code assistance and engineering simulations could look in an updated curriculum.

The paper does well at laying out the motivation: traditional instruction has not kept pace with AI-assisted workflows in industry, and partnerships can help close that distance. The description of embedding the tools is straightforward and could serve as a practical outline for someone redesigning a course.

The main weakness is that the enhancement claims rest entirely on the untested idea that using these tools in class and on the job will produce better programmers. There are no metrics, no comparison to standard teaching, and no discussion of how outcomes would be tracked. It reads as a forward-looking proposal rather than a report on implemented results.

This work is aimed at software engineering faculty who want ideas for modernizing their programs with current AI tools. It offers a template but little else for someone seeking evidence or generalizable findings.

I would not recommend sending it for peer review as a research paper because the absence of any evaluation makes the central assertions impossible to assess. A venue that publishes curriculum proposals might consider it, but it needs more substance to justify referee time.

Referee Report

1 major / 1 minor

Summary. The manuscript describes an innovative collaborative teaching model for software engineering education that integrates Large Language Models (LLMs) and the Model Context Protocol (MCP) into daily teaching activities, code assistance, engineering simulations, and industry internships. It claims that this integration effectively bridges the gap between traditional academic instruction and industrial workflows while enhancing students' programming competence, practical problem-solving abilities, and proficiency with intelligent engineering tools.

Significance. If substantiated with empirical evidence, the described framework could offer a timely pathway for reforming software engineering curricula to better prepare students for AI-augmented industrial practices, addressing the rapid evolution of tools like LLMs in the field.

major comments (1)

Abstract: The assertion that embedding LLM and MCP driven tools 'effectively bridges the gap' and 'enhances students' programming competence, practical problem-solving abilities, and proficiency in using intelligent engineering tools' is presented without any empirical support. No evaluation methods, metrics, sample sizes, baselines, or statistical analyses are provided to demonstrate these enhancements, rendering the central claims declarative rather than evidenced.

minor comments (1)

The manuscript would benefit from defining the Model Context Protocol (MCP) upon first use, as it may not be familiar to all readers in the software engineering education community.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment on the abstract. We agree that the current wording presents claims without empirical backing and will revise the manuscript to address this.

read point-by-point responses

Referee: Abstract: The assertion that embedding LLM and MCP driven tools 'effectively bridges the gap' and 'enhances students' programming competence, practical problem-solving abilities, and proficiency in using intelligent engineering tools' is presented without any empirical support. No evaluation methods, metrics, sample sizes, baselines, or statistical analyses are provided to demonstrate these enhancements, rendering the central claims declarative rather than evidenced.

Authors: We agree that the abstract makes strong declarative claims about effectiveness and enhancements without providing empirical evidence such as evaluation methods, metrics, sample sizes, baselines, or statistical analyses. The manuscript primarily describes the design and rationale of the proposed collaborative teaching model based on classroom and industry observations, rather than reporting results from a formal empirical study. To address this, we will revise the abstract to use more measured language (e.g., 'aims to bridge the gap' and 'is intended to enhance') and add a brief section outlining planned evaluation approaches, metrics, and future validation work. revision: yes

Circularity Check

0 steps flagged

No circularity; purely descriptive framework with no derivations or reductions

full rationale

The paper presents a descriptive teaching model integrating LLMs and MCP into software engineering education, with claims about bridging gaps and enhancing competencies stated declaratively in the abstract and text. No equations, quantitative predictions, fitted parameters, or derivation chains exist. No self-citations, ansatzes, or uniqueness theorems are invoked in a load-bearing manner. The central statements do not reduce to inputs by construction, satisfying the default expectation of no circularity for non-quantitative papers.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no equations, datasets, or explicit modeling choices, so no free parameters, axioms, or invented entities can be identified.

pith-pipeline@v0.9.1-grok · 5703 in / 1050 out tokens · 24858 ms · 2026-06-26T20:19:32.686096+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

13 extracted references · 2 linked inside Pith

[1]

A survey of large language models,

W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Dong, et al., “A survey of large language models,” arXiv preprint arXiv:2303.18223, 2023

Pith/arXiv arXiv 2023
[2]

Introducing the Model Context Protocol,

Anthropic, “Introducing the Model Context Protocol,” Tech. Rep. AN- 2024-1125, Nov. 2024

2024
[3]

A survey of ai agent protocols,

Y . Yang, H. Chai, Y . Song, S. Qi, M. Wen, N. Li, J. Liao, H. Hu, J. Lin, G. Chang, et al., “A survey of ai agent protocols,” arXiv preprint arXiv:2504.16736, 2025

arXiv 2025
[4]

Perspectives on the gap between the software industry and the software engineering education,

D. Oguz and K. Oguz, “Perspectives on the gap between the software industry and the software engineering education,”IEEE Access, vol. 7, pp. 117527–117543, 2019

2019
[5]

Work-based Learning and Universities in the UK: a review of current practice and trends,

E. Foster and J. Stephenson, “Work-based Learning and Universities in the UK: a review of current practice and trends,”Higher Education Research & Development, vol. 17, no. 2, pp. 155–170, 1998

1998
[6]

Students’ Use of GitHub Copilot for Working with Large Code Bases,

A. Shah, A. Chernova, E. Tomson, L. Porter, W. G. Griswold, and A. G. Soosai Raj, “Students’ Use of GitHub Copilot for Working with Large Code Bases,” inProc. 56th ACM Tech. Symp. Comput. Sci. Educ., 2025, pp. 1050–1056

2025
[7]

Do ai assistants help students write formal specifications? a study with chatgpt and the b-method,

A. Capozucca, D. Yampolskyi, A. Goldberg, and M. Cristi ´a, “Do ai assistants help students write formal specifications? a study with chatgpt and the b-method,” in2025 IEEE/ACM 37th Int. Conf. Softw. Eng. Educ. Train. (CSEE&T), 2025, pp. 19–29

2025
[8]

A survey on model context protocol: Architecture, state-of- the-art, challenges and future directions,

P. P. Ray, “A survey on model context protocol: Architecture, state-of- the-art, challenges and future directions,” Authorea Preprints, 2025

2025
[9]

A systematic mapping study on practical approaches to teaching software engineering,

M. R. Marques, A. Quispe, and S. F. Ochoa, “A systematic mapping study on practical approaches to teaching software engineering,” in2014 IEEE Frontiers Educ. Conf. (FIE), 2014, pp. 1–8

2014
[10]

Anderson and B

E. Anderson and B. Schiano,Teaching with cases: A practical guide. Harvard Business Review Press, 2014

2014
[11]

Personalized adaptive learning in higher education: A scoping review of key characteristics and impact on academic performance and engagement,

E. Du Plooy, D. Casteleijn, and D. Franzsen, “Personalized adaptive learning in higher education: A scoping review of key characteristics and impact on academic performance and engagement,”Heliyon, vol. 10, no. 21, 2024

2024
[12]

The development and evaluation of a partial talent management competency model,

A.-M. Oehley, “The development and evaluation of a partial talent management competency model,” Ph.D. dissertation, Stellenbosch Univ., Stellenbosch, South Africa, 2007

2007
[13]

Understanding how enterprises adopt the model context protocol for LLM-driven software engineering,

K. Chen, Y . Sun, J. Keung, Z. Mao, and X. Ma, “Understanding how enterprises adopt the model context protocol for LLM-driven software engineering,” arXiv preprint arXiv:2606.09182, 2026

Pith/arXiv arXiv 2026

[1] [1]

A survey of large language models,

W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Dong, et al., “A survey of large language models,” arXiv preprint arXiv:2303.18223, 2023

Pith/arXiv arXiv 2023

[2] [2]

Introducing the Model Context Protocol,

Anthropic, “Introducing the Model Context Protocol,” Tech. Rep. AN- 2024-1125, Nov. 2024

2024

[3] [3]

A survey of ai agent protocols,

Y . Yang, H. Chai, Y . Song, S. Qi, M. Wen, N. Li, J. Liao, H. Hu, J. Lin, G. Chang, et al., “A survey of ai agent protocols,” arXiv preprint arXiv:2504.16736, 2025

arXiv 2025

[4] [4]

Perspectives on the gap between the software industry and the software engineering education,

D. Oguz and K. Oguz, “Perspectives on the gap between the software industry and the software engineering education,”IEEE Access, vol. 7, pp. 117527–117543, 2019

2019

[5] [5]

Work-based Learning and Universities in the UK: a review of current practice and trends,

E. Foster and J. Stephenson, “Work-based Learning and Universities in the UK: a review of current practice and trends,”Higher Education Research & Development, vol. 17, no. 2, pp. 155–170, 1998

1998

[6] [6]

Students’ Use of GitHub Copilot for Working with Large Code Bases,

A. Shah, A. Chernova, E. Tomson, L. Porter, W. G. Griswold, and A. G. Soosai Raj, “Students’ Use of GitHub Copilot for Working with Large Code Bases,” inProc. 56th ACM Tech. Symp. Comput. Sci. Educ., 2025, pp. 1050–1056

2025

[7] [7]

Do ai assistants help students write formal specifications? a study with chatgpt and the b-method,

A. Capozucca, D. Yampolskyi, A. Goldberg, and M. Cristi ´a, “Do ai assistants help students write formal specifications? a study with chatgpt and the b-method,” in2025 IEEE/ACM 37th Int. Conf. Softw. Eng. Educ. Train. (CSEE&T), 2025, pp. 19–29

2025

[8] [8]

A survey on model context protocol: Architecture, state-of- the-art, challenges and future directions,

P. P. Ray, “A survey on model context protocol: Architecture, state-of- the-art, challenges and future directions,” Authorea Preprints, 2025

2025

[9] [9]

A systematic mapping study on practical approaches to teaching software engineering,

M. R. Marques, A. Quispe, and S. F. Ochoa, “A systematic mapping study on practical approaches to teaching software engineering,” in2014 IEEE Frontiers Educ. Conf. (FIE), 2014, pp. 1–8

2014

[10] [10]

Anderson and B

E. Anderson and B. Schiano,Teaching with cases: A practical guide. Harvard Business Review Press, 2014

2014

[11] [11]

Personalized adaptive learning in higher education: A scoping review of key characteristics and impact on academic performance and engagement,

E. Du Plooy, D. Casteleijn, and D. Franzsen, “Personalized adaptive learning in higher education: A scoping review of key characteristics and impact on academic performance and engagement,”Heliyon, vol. 10, no. 21, 2024

2024

[12] [12]

The development and evaluation of a partial talent management competency model,

A.-M. Oehley, “The development and evaluation of a partial talent management competency model,” Ph.D. dissertation, Stellenbosch Univ., Stellenbosch, South Africa, 2007

2007

[13] [13]

Understanding how enterprises adopt the model context protocol for LLM-driven software engineering,

K. Chen, Y . Sun, J. Keung, Z. Mao, and X. Ma, “Understanding how enterprises adopt the model context protocol for LLM-driven software engineering,” arXiv preprint arXiv:2606.09182, 2026

Pith/arXiv arXiv 2026