Browse categories
Explore
Fiverr Pro
English
$
USD
I will build deep learning, nlp and computer vision models in python
I will design and develop your website, mobile app with AI automation
I'm a UI/UX designer and web developer who turns ideas into high-performing digital products. From mobile app design and WordPress development to landing pages and dashboards, I craft pixel-perfect in...
Level 1
Has met certain performance criteria and shows strong potential in the marketplace.
About this Gig
From image classifiers to text models, I build deep learning, NLP and computer vision models in Python that are trained, evaluated and documented properly.
I am K. M. Arafat Islam, Founder and CEO of BuildSign Digital, CSE graduate with 4 years of experience, 15 published papers in Springer and IEEE venues and a government-certified app developer (EDGE).
Computer vision: Image classification with CNNs and transfer learning (ResNet, EfficientNet, ViT) Object detection with YOLO Image segmentation (U-Net)
NLP: Text classification and sentiment analysis Fine-tuning BERT and transformer models Named entity recognition and topic modeling
Also: LSTM and GRU for time series.
You receive: Commented notebook (PyTorch or TensorFlow/Keras) Trained model weights Accuracy, F1, loss curves and confusion matrix Short technical report
I build and document models on your data; I do not write research content on your behalf. Message me your dataset and task first.
My Portfolio
Other Data Science & ML Services I Offer
FAQ
Do you provide GPUs?
I train on Google Colab or Kaggle GPUs for most projects. For very large training runs, we can use your cloud GPU or budget for compute.
Can you fine-tune BERT or other transformers?
Yes. I fine-tune BERT, RoBERTa, DistilBERT and similar models with Hugging Face for classification, NER and sentiment tasks.
Which vision models do you use?
ResNet, EfficientNet, MobileNet, ViT, YOLO and U-Net, chosen to fit your data size and accuracy needs.
Do I need labeled data?
For supervised tasks, yes. I can advise on labeling formats and small-data strategies such as transfer learning and augmentation.
Can you explain the model's decisions?
Yes. Grad-CAM and attention maps are available as extras, and SHAP or LIME through my explainable AI gig.
Can you promise a target accuracy?
Not before seeing the data. I report honest, validated results and explain the limits.
