Browse categories
Explore
Fiverr Pro
English
$
USD
I will build computer vision and image classification models
AI and ML Engineer
I am a Machine Learning Engineer with experience developing machine learning, deep learning, computer vision, and natural language processing (NLP) solutions using Python. I build reliable models for ...
About this Gig
Are you looking for custom Computer Vision or Deep Learning solutions built in PyTorch?
I specialize in building, fine-tuning, and evaluating production-ready vision models for real-world applications.
What I Can Build For You:
- Image Classification & Medical AI: CNNs, MobileNetV2, ResNet, EfficientNet.
- Object Detection: Custom YOLO models, real-time tracking, and OpenCV integration.
- Model Tuning & Optimization: Class-imbalance handling, custom loss functions, and threshold optimization (maximizing Recall/Precision).
- Explainable AI (XAI): Spatial Grad-CAM heatmaps to visually verify model predictions.
- Custom Inference APIs: FastAPI microservices to deploy your models.
What You Will Receive:
- Well-documented Google Colab notebook (.ipynb) or modular Python source code (.py).
- Trained model weights (.pt / .pth) and configuration files.
- Full evaluation suite: Seaborn confusion matrices, classification reports, and training curves.
Please contact me before placing an order to discuss your dataset and model requirements!
APIs:
Other
Programming language:
Python
•
Colab
Tools:
Jupyter Notebook
•
OpenCV
•
TensorFlow
•
Excel
•
Colab
•
PyTorch
Frameworks:
Scikit-learn
•
Keras
•
PyTorch
•
Panda
My Portfolio
Other Data Science & ML Services I Offer
FAQ
What do I need to provide before placing an order?
Please provide a clear summary of your project goals and access to your dataset (or a sample subset) via Google Drive, Kaggle, or Zip. If you don't have a dataset yet, contact me first and I can help you source or structure one.
Will I receive the complete Python code and trained model files?
Yes! Every package includes the full, fully-documented Google Colab notebook (.ipynb) or clean Python source scripts (.py), along with the trained PyTorch weights (.pt or .pth) and configuration files.
Can you handle small, highly imbalanced datasets?
Absolutely. I utilize specialized techniques such as class weighting (pos_weight), focal loss, precision-recall curve analysis, and targeted decision threshold optimization to ensure strong model performance even on severely imbalanced data.
How do I know the model isn't just a "black box" guessing predictions?
I provide spatial Grad-CAM heatmaps that visually highlight exact pixel regions driving the model's decisions, alongside full Seaborn confusion matrices and classification metrics (Precision, Recall, F1-Score).
Can you deploy the model as an API for my web or mobile app?
Yes, under the Premium package (or as a custom add-on), I can build a lightweight FastAPI microservice to serve predictions directly to your application or web backend.
