You will get expert image classification and model development services
Top Rated
Project details
You will get professional image classification services powered by deep learning and advanced computer vision expertise. With over 5 years of AI engineering experience, I specialize in building accurate classification models using frameworks such as PyTorch and TensorFlow.
From dataset preparation and annotation to model development and deployment, I deliver end-to-end solutions optimized for your business needs. Whether you require binary, multi-class, or fine-grained classification, I ensure high-performance results with rigorous validation, documentation, and production-ready code.
From dataset preparation and annotation to model development and deployment, I deliver end-to-end solutions optimized for your business needs. Whether you require binary, multi-class, or fine-grained classification, I ensure high-performance results with rigorous validation, documentation, and production-ready code.
Machine Learning Tools
Keras, NumPy, OpenCV, Python, Python Scikit-Learn, PyTorch, scikit-learn, SciPy, TensorFlowWhat's included
| Service Tiers |
Starter
$150
|
Standard
$450
|
Advanced
$800
|
|---|---|---|---|
| Delivery Time | 5 days | 10 days | 15 days |
Number of Revisions | 2 | 3 | 5 |
Number of Model Variations | 2 | 5 | 7 |
Number of Scenarios | 1 | 3 | 5 |
Number of Graphs/Charts | 2 | 4 | 6 |
Model Validation/Testing | - | ||
Model Documentation | - | ||
Data Source Connectivity | - | - | |
Source Code | - | - |
Frequently asked questions
6 reviews
(5)
(1)
(0)
(0)
(0)
This project doesn't have any reviews.
DG
Dhruv G.
May 12, 2026
Computer Vision | AI | Mentoring | Python DSA
Thanks a lot shah, really appreciate the efforts for educating me, and sharing your knowledge woth me. Really appreciate the efforts, thanks for all thr lessons on python, ai, ml, dl, python and concepts foundations and lessons amd practice.
LP
Ljiljana P.
Apr 23, 2026
Looking for AI, ML Engineer
BB
Burim B.
Feb 21, 2026
Object Detection Research Project: Data Collection, YOLO Training, and Research Paper Writing
I received the support I needed in a clear and professional manner, which made the entire process smooth and stress-free.
RM
Ray M.
Feb 11, 2026
Ocr
He is genuine and honest, and very flexible with time and goal focused, will work with him again
ID
Imran D.
Aug 12, 2025
Build High-Accuracy Object Detection + Identity Tracking Pipeline for Multi-Person Video Clips
About Shah
Computer Vision Engineer | YOLO | Tracking | OCR | Edge AI | VLM
92%
Job Success
Wah Cantt, Pakistan - 5:23 am local time
I'm a Computer Vision Engineer with a proven track record of shipping real-world CV systems that perform under production conditions. From multi-object tracking with ReID across scene changes, to industrial OCR pipelines for logistics, to real-time product detection for retail automation — I've built systems that clients trust and deploy.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
WHAT I BUILD
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Object Detection & Custom Model Training
Custom-trained YOLO (v8–v11) and RT-DETR models on your specific dataset — with data collection, annotation, augmentation, training, evaluation, and deployment-ready exports (ONNX, TensorRT).
Multi-Object Tracking & Re-Identification
Production-grade tracking pipelines with ReID that maintain identity across occlusions, camera cuts, and scene changes — using ByteTrack, DeepSORT, and custom embedding models.
OCR & Document Intelligence
End-to-end OCR pipelines using PaddleOCR, EasyOCR, and Tesseract — including container/label recognition for logistics, display reading, and structured data extraction from images.
Visual Inspection & Industrial Automation
Real-time defect detection, anomaly detection, and quality control systems — built for manufacturing, logistics, retail, and healthcare environments.
Edge AI Deployment
Optimized model deployment on NVIDIA Jetson and Raspberry Pi using TensorRT, ONNX, and DeepStream — achieving real-time inference at the edge.
Visual Language Models (VLM)
Multimodal vision-language pipelines using AWS Nova models hosted on Amazon Bedrock — for image understanding, visual Q&A, and scene description at scale.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
REAL PROJECTS DELIVERED
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
→ Multi-person detection & tracking pipeline with cross-scene ReID for video analysis
→ Real-time smart bar system detecting customers and products simultaneously
→ Container OCR validation system for logistics — reading, parsing, and verifying codes
→ Treadmill monitor using YOLO + OCR to read fitness machine displays
→ Real-time glove detection system using YOLO object detection
→ Custom YOLO model training for research with published research paper
→ Multimodal VLM pipeline using AWS Nova on Amazon Bedrock
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
TECH STACK
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✅ Detection: YOLOv8–v11 · RT-DETR · SAM · Faster R-CNN
✅ Tracking: ByteTrack · DeepSORT · StrongSORT · ReID
✅ OCR: PaddleOCR · EasyOCR · Tesseract · AWS Textract
✅ VLM: AWS Nova · Amazon Bedrock · Multimodal Pipelines
✅ Vision: OpenCV · Albumentations · Supervision · Roboflow
✅ Frameworks: PyTorch · TensorFlow · ONNX · TensorRT
✅ Deployment: FastAPI · Docker · AWS · NVIDIA Jetson · Raspberry Pi
✅ Annotation: CVAT · LabelImg · Roboflow · Label Studio
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
WHY CLIENTS CHOOSE ME
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
→ 5 CV projects in portfolio with real-world deployment results
→ 100% 5-star ratings on all Computer Vision jobs
→ Top Rated on Upwork · Fast delivery · Clear communication
→ I train on YOUR data, not generic demos
→ Full pipeline: data → model → API → edge device
Have a Computer Vision challenge? Let's solve it. Message me today.
Steps for completing your project
After purchasing the project, send requirements so Shah can start the project.
Delivery time starts when Shah receives requirements from you.
Shah works on your project following the steps below.
Revisions may occur after the delivery date.
Requirement Gathering
Understand your use case, dataset, and project goals.
Data Preparation
Clean, preprocess, and (if needed) annotate the dataset.