Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -10,4 +10,44 @@ pinned: false
|
|
| 10 |
license: mit
|
| 11 |
---
|
| 12 |
|
| 13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
license: mit
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# π§ Resume Screener Agent - AI-Powered Matching Tool
|
| 14 |
+
|
| 15 |
+
This Space hosts a LangChain-powered Resume Screening Agent that matches resumes to job descriptions with structured, explainable scoring. Designed to run on CPU and lightweight models, it works perfectly within Hugging Face Spaces.
|
| 16 |
+
|
| 17 |
+
---
|
| 18 |
+
|
| 19 |
+
## π Features
|
| 20 |
+
|
| 21 |
+
- β
Match resumes with job descriptions in seconds
|
| 22 |
+
- β
Score compatibility (0β100)
|
| 23 |
+
- β
Extract matched skills
|
| 24 |
+
- β
Explain decisions with structured JSON output
|
| 25 |
+
- β
Runs entirely on CPU (swap OpenAI with HuggingFacePipeline for full offline)
|
| 26 |
+
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
## π οΈ How It Works
|
| 30 |
+
|
| 31 |
+
1. Paste plain text **resume** and **job description**.
|
| 32 |
+
2. Agent evaluates using an LLM prompt chain.
|
| 33 |
+
3. Outputs:
|
| 34 |
+
- `MatchScore` (0β100)
|
| 35 |
+
- `KeySkills` (list)
|
| 36 |
+
- `Justification` (text)
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
+
|
| 40 |
+
## π‘ Built With
|
| 41 |
+
|
| 42 |
+
- [LangChain](https://www.langchain.com/)
|
| 43 |
+
- [Gradio](https://gradio.app/)
|
| 44 |
+
- CPU-compatible LLMs (e.g., `flan-t5-base`, `oasst-sft-1-pythia`)
|
| 45 |
+
|
| 46 |
+
---
|
| 47 |
+
|
| 48 |
+
## π¦ Requirements
|
| 49 |
+
|
| 50 |
+
```txt
|
| 51 |
+
gradio
|
| 52 |
+
langchain
|
| 53 |
+
openai # Optional, replace with transformers + HuggingFacePipeline
|