# Talent-Match-AI: Resume and Job Description Matching

## 📌 Overview

This repository hosts the quantized version of the **BERT-base-uncased** model for **Resume and Job Description Matching**. The model is designed to determine whether a resume aligns well with a given job description. If they are a strong match, the model outputs "Good Fit" with a confidence score; otherwise, it categorizes them as "Potential Fit" or "Not a Good Fit." The model has been optimized for efficient deployment while maintaining reasonable accuracy, making it suitable for real-time applications.

## 🏰 Model Details

- **Model Architecture:** BERT-base-uncased  
- **Task:** Resume and Job Description Matching  
- **Dataset:** `facehuggerapoorv/resume-jd-match`  
- **Quantization:** Float16 (FP16) for optimized inference  
- **Fine-tuning Framework:** Hugging Face Transformers  

## 🚀 Usage

### Installation

```bash
pip install transformers torch
```

### Loading the Model

```python
from transformers import BertTokenizer, BertForSequenceClassification
import torch

device = "cuda" if torch.cuda.is_available() else "cpu"

model_name = "AventIQ-AI/bert-talentmatchai"
model = BertForSequenceClassification.from_pretrained(model_name).to(device)
tokenizer = BertTokenizer.from_pretrained(model_name)
```

### Resume Matching Inference

```python
import torch

# Set device (use GPU if available)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# Define label mapping
label_mapping = {0: "Not a Good Fit", 1: "Potential Fit", 2: "Good Fit"}

# Sample resume text for testing
test_resume = ["I have worked in different industries and have a lot of experience. I am a hard worker and can learn anything."]

# Tokenize test data
test_tokens = tokenizer(test_resume, padding="max_length", truncation=True, return_tensors="pt").to(device)  # Move input to same device as model

# Make predictions
with torch.no_grad():  # Disable gradient computation for inference
    output = model(**test_tokens)

# Get predicted label
predicted_label = output.logits.argmax(dim=1).item()

# Print result
print(f"Predicted Category: {predicted_label} ({label_mapping[predicted_label]})")

label_mapping = {0: "No Fit", 1: "Low Fit", 2: "Potential Fit", 3: "Good Fit"}
print(f"Predicted Category: {label_mapping[predictions]}")

```

## 📊 Quantized Model Evaluation Results

### 🔥 Evaluation Metrics 🔥

- ✅ **Accuracy:**  0.9224  
- ✅ **Precision:** 0.9212  
- ✅ **Recall:**    0.8450  
- ✅ **F1-score:**  0.7718  

## ⚡ Quantization Details

Post-training quantization was applied using PyTorch's built-in quantization framework. The model was quantized to Float16 (FP16) to reduce model size and improve inference efficiency while balancing accuracy.

## 💽 Repository Structure

```
.
├── model/               # Contains the quantized model files
├── tokenizer_config/    # Tokenizer configuration and vocabulary files
├── model.safetensors/   # Quantized Model
├── README.md            # Model documentation
```

## ⚠️ Limitations

- The model may struggle with resumes and job descriptions that use non-standard terminology.
- Quantization may lead to slight degradation in accuracy compared to full-precision models.
- Performance may vary across different industries and job levels.

## 🤝 Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request if you have suggestions or improvements.