I've been using Iris.ai for my cloud certification studies, mainly to find papers on specific AWS service architectures. The suggestions are good, but I'm a junior engineer, so I needed a way to check if the papers it finds are actually relevant and not over my head.
Here's my simple validation workflow. I use the API to fetch suggestions, then cross-check with my own notes.
I wrote a small Python script that calls the Iris.ai API (using the "focus" endpoint) and filters results based on my own keyword list. I store my known-good sources (like AWS Whitepapers, specific conference names) in a config file.
```python
# Simplified example - I filter by keywords in the title/abstract
my_keywords = ["aws lambda", "serverless", "fargate", "step functions"]
irrelevant_terms = ["quantum", "genomic", "phylogenetic"] # I get these sometimes!
for paper in iris_suggestions:
title_lower = paper['title'].lower()
if any(term in title_lower for term in irrelevant_terms):
print(f"Skipping: {paper['title']}")
continue
if any(term in title_lower for term in my_keywords):
print(f"Relevant: {paper['title']}")
# I then manually check the abstract against my notes
```
I then manually scan the abstracts of the filtered papers. If more than two key concepts aren't in my personal study notes, I tag it as "review later" when I'm more advanced. This keeps my reading list focused. Does anyone else have a method for this? Maybe using tags or something in Zotero?