I will build an end to end ai system in computer vision, medical ai nlp generative ai
About this gig
Looking for AI that actually works not just a notebook?
I'm Samama, an Applied AI Engineer (MS AI, 4.0 CGPA | BEng Computer Engineering). I build production-grade systems across Computer Vision, Medical AI, NLP, and Generative AI.
What I've shipped:
MedIntel ViT chest X-ray classifier, Grad-CAM, FHIR export, 83.5% accuracy ️
Satellite segmentation U-Net, 79.7% accuracy, 99% less annotation effort
NLP ensemble classifier SVM, BERT, RF with full evaluation
Federated Learning research work and publications
I build:
CV pipelines classification, detection, segmentation
Medical AI explainable diagnostics, DICOM, FHIR
NLP transformers, fine-tuning, text classification
GenAI RAG, LangChain, LLM apps, chatbots
Every delivery includes:
Clean, documented Python code Evaluation metrics + error analysis README + requirements.txt Gradio/FastAPI interface
Research-backed. Engineer-executed. Zero ghosting.
Let's build something that works. Message me.
Get to know Samama
Applied AI Engineer Computer Vision Medical AI NLP GenAI
- FromPakistan
- Member sinceOct 2015
- Avg. response time1 hour
Languages
English
FAQ
I have my own dataset — can you work with it?
Absolutely. Share your data and requirements upfront and I'll handle everything from preprocessing to final delivery.
Can you build explainable AI systems?
Yes — Grad-CAM, attention rollout, and SHAP are standard in my CV and medical AI deliverables.
Do you handle medical data like DICOM or FHIR?
Yes. I've built DICOM preprocessing pipelines and FHIR-compliant export systems in production-grade medical AI projects.
Can you fine-tune LLMs or build RAG systems?
Yes — LangChain, HuggingFace Transformers, OpenAI API, and vector database integration (FAISS, Chroma).
Can you handle multilingual or non-English text?
Yes — multilingual transformers (mBERT, XLM-R) for Arabic, Urdu, and other languages.
Will the code be clean and reproducible?
Always. Modular, commented code with requirements.txt, README, and clear project structure — every time.

