Assignment 1 submission

[Mandar-Narkar]

What changes are you trying to make? (e.g. Adding or removing code, refactoring existing code, adding reports)**

Add my completed solution for Assignment 1. I implemented a notebook that loads the GenAI Divide report PDF, generates a structured summary using the OpenAI SDK and Pydantic, and evaluates the summary with multiple DeepEval metrics.

What did you learn from the changes you have made?

I learned how to separate secrets from code with dotenv, how to build a structured output schema with Pydantic, and how to call the OpenAI client to produce summaries in a specific tone. I also learned how summarisation metrics like coverage, alignment, coherence, tone, and safety can be implemented and interpreted.

Was there another approach you were thinking about making? If so, what approach(es) were you thinking of?

I considered using a web article instead of the PDF and using a simpler evaluation (only one summarisation score), but chose a longer PDF plus multiple metrics to better test model behaviour.

Were there any challenges? If so, what issue(s) did you face? How did you overcome it?

One challenge was handling the 26‑page PDF and ensuring all page content was concatenated correctly before sending it to the model. Another was avoiding metric reuse in DeepEval, which could cache scores. I addressed these by joining page contents into a single document_text string and writing a factory function that returns fresh metric instances for each evaluation run.

How were these changes tested?

I ran the notebook end‑to‑end, verified that the summary and evaluation results were produced without errors, and confirmed that scores and reasons were updated when changing the summary generation prompt.

A reference to a related issue in your repository (if applicable)

N/A (course assignment).

Checklist

• I can confirm that my changes are working as intended