Skip to content

Commit a2e4190

Browse files
committed
Week 3 updates by Aman Naik
1 parent 60e5e0e commit a2e4190

File tree

1 file changed

+69
-0
lines changed

1 file changed

+69
-0
lines changed
Lines changed: 69 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,69 @@
1+
---
2+
title: "DMP ’25 Week 3 Update by Aman Naik"
3+
excerpt: "This week's focus was developing a working demo for the Story Builder feature using Streamlit and gathering mentor feedback for further improvements."
4+
category: "DEVELOPER NEWS"
5+
date: "2025-06-21"
6+
slug: "2025-06-21-dmp-25-AmanNaik-week03"
7+
author: "@/constants/MarkdownFiles/authors/amannaik247.md"
8+
tags: "dmp25,writeactivity,write,sugarlabs,week03,amannaik247"
9+
image: "assets/Images/c4gt_DMP.png"
10+
---
11+
12+
<!-- markdownlint-disable -->
13+
14+
# Week 3 Progress Report by Aman Naik
15+
16+
**Project:** [Add an AI-assistant to the Write Activity](https://github.com/sugarlabs/write-activity/issues/52)
17+
**Mentors:** [Walter Bender](https://github.com/walterbender), [Ibiam Chihurumnaya](https://github.com/chimosky)
18+
**Reporting Period:** 2025-06-16 – 2025-06-21
19+
20+
---
21+
22+
## Goals for This Week
23+
24+
- **Goal 1:** Develop a working demo of the Story Builder feature
25+
- **Goal 2:** Implement the demo using Streamlit
26+
- **Goal 3:** Gather feedback from mentors and plan improvements
27+
28+
---
29+
30+
## This Week’s Achievements
31+
32+
1. **Developed a Demo of the Story Builder Feature**
33+
- Created a functional web application using Streamlit to demonstrate how the story builder AI guides students through storytelling. The demo mimics the question-and-answer-based flow and simulates an AI companion.
34+
- Find the demo [here](https://story-builder-ai.streamlit.app/)
35+
36+
2. **Presented the Demo to Mentors**
37+
- Shared the working version with my mentors. I received positive feedback and valuable suggestions on what could be improved and how the feature can be enhanced with new ideas.
38+
39+
---
40+
41+
## Challenges & How I Overcame Them
42+
43+
- **Challenge:** Designing a feedback loop where the AI accumulates all context from the student’s responses
44+
**Solution:** After repeated iterations and research, I realized that assigning a distinct personality to the LLM helps it collect and remember context more naturally. This personality-driven approach ensures a more engaging and coherent interaction throughout the story-building process.
45+
46+
---
47+
48+
## Key Learnings
49+
50+
**Built and Deployed a Streamlit Demo to Simulate Story Interaction**
51+
- Gained practical experience in using Streamlit to build and deploy web apps that demonstrate AI interaction flow for educational use cases.
52+
53+
**Learned the Importance of AI Personality in Context Retention**
54+
- Discovered that crafting an AI assistant with a personality improves its ability to retain context and makes the storytelling experience more relatable and fun for children.
55+
56+
---
57+
58+
## Next Week’s Roadmap
59+
60+
- Begin fine-tuning a language model using AWS
61+
- Build a system where the AI recommends books similar to the student's story at the end of the conversation
62+
63+
---
64+
65+
## Acknowledgments
66+
67+
Thank you to my mentors, the Sugar Labs community, and fellow contributors for their continuous guidance and valuable feedback!
68+
69+
---

0 commit comments

Comments
 (0)