I will build and deploy custom machine learning and computer vision models


About this gig
Need a machine learning or computer vision model built and actually working in production not just a Jupyter notebook?
I'm a full-stack ML engineer who handles your entire pipeline myself data, model training, and deployment so you're not hiring a data scientist, a backend developer, and a DevOps person separately for one project.
Why work with me:
- One engineer, full pipeline: from model training to a live, deployed API, no handoffs between specialists
- Production-grade: I build real-time computer vision systems for industrial clients daily, not just demos
- Fast turnaround: most builds delivered in 3-10 days depending on complexity
What I build:
- Computer vision models (object detection, classification, image/video analytics)
- Neural networks and deep learning models (CNNs, transfer learning)
- Predictive ML models (classification, regression, forecasting)
- Deployment-ready delivery as a live API (FastAPI) or app (Streamlit)
You'll get:
- Clean source code + setup guide
- Trained model with accuracy/performance metrics
- Tested API or demo with real output examples
- Optional deployment support
Message me before ordering with your dataset, goals and any sample file.
Get to know Tashfeen R
AI and ML Engineer, LLM Agents and Automation
- FromPakistan
- Member sinceJan 2025
Languages
Urdu, English, German
My Portfolio
FAQ
What do you need from me to start?
Dataset (CSV/Excel/JSON), target column, problem goal, and any constraints/deadline.
Do you work on classification and regression?
Yes both. I will recommend the right approach based on your data and objective.
Will you provide evaluation and a report?
Yes. I include metrics and error analysis to ensure the model is reliable.
Can you handle imbalanced datasets?
Yes class weights, sampling strategies, threshold tuning, and proper PR/ROC evaluation.
Do you provide source code?
Yes, clean and documented source code is included.
Can you deploy on cloud (AWS/GCP/Azure)?
Currently I provide local/Docker-ready deliverables. Cloud deployment can be added later.

