Draft - Paper: From Hello World to Hello LLM: A Python Developer’s Survival Guide

[curiousjazz77]

This paper shares practical lessons from building AI-powered tools for educational institutions during a social impact hackathon. Within it, we cover LLM patterns in Python including

• function calling
• prompt constraints
• agent loops
• RAG

Each example is accompanied by code snippets and guidance on when to apply each.

@audrey-webb

[github-actions]

Curvenote Preview

Directory	Preview	Checks	Updated (UTC)
papers/from-hello-world-to-hello-llm-a-python-developers-survival-guide	🔍 Inspect	✅ 23 checks passed (4 optional)	Aug 15, 2026, 4:12 PM

[JennEYoon]

Dear Jasmine @curiousjazz77 and Audrey (@audrey-webb),

I am Jennifer Yoon (@JennEYoon), co-chair of the SciPy 2026 Proceedings Committee. I am honored to serve as the Editor for this paper. Please feel free to reach out to me with any questions about the Proceedings review process.

Reviewers, please introduce yourselves here.

Important upcoming dates:

• July 25: Reviewers — First draft full reviews due
• August 22: Reviewers — Final draft full reviews due
• August 30: Authors — Final submission date
• September 5: Reviewers — Final up/down vote due

Authors, please try to respond to all substantive reviewer comments during each round of reviews. Feel free to ask the reviewers for clarification. Because this is an iterative process, please check in every few days to see if there are any open issues. And feel free to revise your paper at any time up until the final submission date.

[JennEYoon]

Reviewers, here are additional information.

Reviewers -- readme: https://github.com/scipy-conference/scipy_proceedings/blob/2026/README.md#instructions-for-reviewers

Reviewers -- How to add comments on GitHub Pull Request:
https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/reviewing-changes-in-pull-requests/commenting-on-a-pull-request

[decorouz]

Dear authors,

I am Biola Adeyemi. I am volunteering to review this paper for the Scipy 2026 Proceedings

[JennEYoon]

Hi Biola, Thanks for accepting!

[JennEYoon]

Dear Niharika (@niharika13pjr),
Thank you for agreeing to review this paper!
Jennifer

[niharika13pjr]

This is a useful and timely proposal that gives Python developers a practical introduction to building AI applications without requiring a machine learning background. I particularly liked the focus on treating AI as another system component that requires validation, monitoring, and clear guardrails.

However, I found that some of the code examples do not fully demonstrate the concepts described in the text. For example, the RAG example uses a keyword check and a direct SQL query but does not show embeddings or semantic retrieval. The agent example also follows a fixed sequence of function calls rather than allowing the system to choose its next step based on earlier results.

There is also a mismatch in the function-calling example. The user asks which individual students are most at risk, but the function only returns an average risk score for the school. This should also be reviewed against the earlier requirement that only aggregate data should be returned.

I recommend revising the examples so that they more clearly represent each pattern and follow a consistent approach to student data privacy. One complete end-to-end example showing the user question, retrieval or tool execution, validation, and final response would make the presentation easier to follow and more useful for attendees.

[niharika13pjr]

Thank you for the updates to the paper. I reviewed the current version again with my earlier comments in mind. The paper provides a useful introduction to common AI development patterns in Python, but I still have a few technical concerns with the examples.

• RAG example: The current example uses a keyword check followed by a SQL query. This does not fully match the discussion of embeddings and retrieval pipelines. I suggest either adding a simple retrieval example or describing this as database-backed context retrieval rather than RAG.

• Agent example: The steps are executed in a fixed order in Python, and the LLM only summarizes the final result. This looks more like a multi-step pipeline than an agent loop. I suggest either adjusting the terminology or showing how the model decides which tool to call based on intermediate results.

• Function-calling example: The user asks which students are most at risk, but the function returns only an average risk value for the school. Please align the question with the data returned by the function.

• Privacy consistency: Earlier in the paper, the system is described as returning only aggregate data and not exposing individual student information. Later examples ask which individual students are at risk. I recommend making these examples consistent with the privacy rules described earlier.

These issues should not require major changes. In several places, updating the wording, example question, or terminology would be enough. Addressing them would make the examples clearer and more technically accurate for readers.

GenAI disclosure: I used ChatGPT to help refine the wording and clarity of my reviewer-authored comments. The technical assessment and recommendations are my own, and I reviewed the final text before submission.

[JennEYoon]

Thank you @niharika13pjr for your full review including the GenAI disclosure.

Authors @curiousjazz77 and @audrey-webb, please respond to this review within the coming week.

[curiousjazz77]

Thank you @niharika13pjr for your full review including the GenAI disclosure.

Authors @curiousjazz77 and @audrey-webb, please respond to this review within the coming week.

Thank you all for your feedback! We truly appreciate it. As requested, we've made a few updates for your review.