Skip to content

Commit bf86793

Browse files
DMP '25 Week09 update by Harshit Verma (sugarlabs#364)
1 parent befebd7 commit bf86793

File tree

1 file changed

+78
-0
lines changed

1 file changed

+78
-0
lines changed
Lines changed: 78 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,78 @@
1+
---
2+
title: "DMP ’25 Week 09 Update by Harshit Verma"
3+
excerpt: "Finalized the /debug endpoint in Sugar-AI and tested its performance using buggy Python code. I also plan to experimented with a larger model to evaluate debugging responses and began refining them to be more age-appropriate."
4+
category: "DEVELOPER NEWS"
5+
date: "2025-08-09"
6+
slug: "2025-08-09-dmp-25-therealharshit-week09"
7+
author: "@/constants/MarkdownFiles/authors/harshit-verma.md"
8+
tags: "dmp25,sugarlabs,week08,therealharshit"
9+
image: "assets/Images/c4gt_DMP.png"
10+
---
11+
12+
<!-- markdownlint-disable -->
13+
14+
# Week 09 Progress Report by Harshit Verma
15+
16+
**Project:** [LLM-powered Debugger for Pippy](https://github.com/sugarlabs/Pippy/issues/95)
17+
**Mentors:** [Walter Bender](https://github.com/walterbender), [Ibiam Chihurumnaya](https://github.com/chimosky), [Kshitij Shah](https://github.com/kshitijdshah99)
18+
**Assisting Mentors:** [Devin Ulibarri](https://github.com/pikurasa)
19+
**Reporting Period:** 2025-07-28 - 2025-08-03
20+
21+
---
22+
23+
## Goals for This Week
24+
25+
- **Goal 1:** Complete work on the `/debug` endpoint in Sugar-AI.
26+
- **Goal 2:** Test the debugging response using Sugar-AI on various buggy Python code samples.
27+
28+
---
29+
30+
## This Week’s Achievements
31+
32+
1. **Enhanced `/debug` Endpoint in Sugar-AI**
33+
- Finalized the logic for better parsing and structured debugging feedback.
34+
- Ensured the endpoint outputs clear, concise, and kid-friendly messages.
35+
36+
2. **Testing with Real Buggy Python Code**
37+
- Curated multiple buggy Python scripts ranging from beginner-level mistakes to more advanced logical errors.
38+
- Verified that Sugar-AI correctly identifies, explains, and suggests fixes for each case.
39+
- Helped improve prompt handling to maintain age-appropriate explanations for kids.
40+
- Document: [LLM response](https://docs.google.com/document/d/1a8GgsccWm9lSuFCWc7yStXzZ3dZW_EIigLyGXseOfjs/edit?usp=sharing)
41+
42+
---
43+
44+
## Challenges & How I Overcame Them
45+
46+
- **Challenge:** Initial debugging responses were too technical for the target audience.
47+
**Solution:** Iterated the prompt design and response formatting to make them simpler and more engaging.
48+
49+
---
50+
51+
## Key Learnings
52+
53+
- Deepened understanding of API endpoint design.
54+
- Improved prompt engineering for educational AI tools.
55+
56+
---
57+
58+
## Next Week’s Roadmap
59+
60+
- Observed that the current responses were too technical, so worked on making them more age-appropriate.
61+
- Tested the debugging responses using a larger language model for improved accuracy.
62+
63+
---
64+
65+
## Resources & References
66+
67+
**Repository**
68+
- [Pippy](https://github.com/therealharshit/Pippy/tree/DMP2025/Pippy-Debugger)
69+
- [sugar-ai](https://github.com/sugarlabs/sugar-ai)
70+
- [pippy-debugger-server](https://github.com/therealharshit/pippy-debugger-server)
71+
72+
---
73+
74+
## Acknowledgments
75+
76+
Thank you to my mentors, the Sugar Labs community, and fellow contributors for your continued guidance and support!
77+
78+
---

0 commit comments

Comments
 (0)