You will get Model conversion + optimization (YOLO & CV) for real-time edge AI
Top Rated
Top Rated
Project details
You will get end-to-end AI model conversion and optimization that makes your computer vision models faster, lighter, and production-ready.
I support multiple platforms, TFLite (Android), CoreML (iOS), TensorRT (NVIDIA), RKNN (Rockchip), and ONNX, ensuring your models run smoothly on mobile or edge devices.
I go beyond simple conversion by applying advanced optimization techniques:
FP32 ➝ FP16/INT8 quantization,
structured & unstructured pruning,
layer fusion, and compression.
This results in up to 5x speed boosts without accuracy loss.
For example,
I optimized YOLOv5s to achieve 70ms per frame on a mid-range Android phone, enabling real-time detection.
On NVIDIA hardware, my TensorRT pipeline delivered a jump from 35 FPS to 120 FPS on dual video streams, showcasing true high-performance edge AI.
With proven expertise in computer vision and deployment, I deliver models that are optimized, benchmarked, and ready for integration. The work I provide is reliable, tested, and high quality. giving you confidence to take your AI project from concept to real-world deployment.
I support multiple platforms, TFLite (Android), CoreML (iOS), TensorRT (NVIDIA), RKNN (Rockchip), and ONNX, ensuring your models run smoothly on mobile or edge devices.
I go beyond simple conversion by applying advanced optimization techniques:
FP32 ➝ FP16/INT8 quantization,
structured & unstructured pruning,
layer fusion, and compression.
This results in up to 5x speed boosts without accuracy loss.
For example,
I optimized YOLOv5s to achieve 70ms per frame on a mid-range Android phone, enabling real-time detection.
On NVIDIA hardware, my TensorRT pipeline delivered a jump from 35 FPS to 120 FPS on dual video streams, showcasing true high-performance edge AI.
With proven expertise in computer vision and deployment, I deliver models that are optimized, benchmarked, and ready for integration. The work I provide is reliable, tested, and high quality. giving you confidence to take your AI project from concept to real-world deployment.
Machine Learning Tools
Amazon SageMaker, Azure Machine Learning, ChatGPT, GitHub Copilot, Google AutoML, Keras, MLflow, NumPy, NVIDIA AI Platform, Open Neural Network Exchange, OpenCV, pandas, Python, Python Scikit-Learn, PyTorch, scikit-learn, TensorFlow, Tesseract OCRWhat's included
| Service Tiers |
Starter
$99
|
Standard
$350
|
Advanced
$999
|
|---|---|---|---|
| Delivery Time | 1 day | 3 days | 5 days |
Number of Revisions | 1 | 2 | Unlimited |
Number of Model Variations | 1 | 1 | 1 |
Number of Scenarios | 1 | 1 | 1 |
Number of Graphs/Charts | 1 | 3 | 3 |
Model Validation/Testing | |||
Model Documentation | - | ||
Data Source Connectivity | - | - | |
Source Code | - | - |
Frequently asked questions
22 reviews
(22)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
EQ
Emil Q.
Jun 3, 2026
Golf Swing Tracer & Club Detection Technical Cons
Muhammad did excellent work on a technically demanding project and was a genuinely easy collaborator.
He was fast, detail-oriented, and clearly experienced. He took the time to properly understand our references and existing logic. Communication was crisp: he asked clarifying questions instead of guessing, kept calls short and focused, and delivered a clear, structured plan without fluff or generic AI-style language.
Overall, I’m very happy with the outcome and would really recommend working with Muhammad.
He was fast, detail-oriented, and clearly experienced. He took the time to properly understand our references and existing logic. Communication was crisp: he asked clarifying questions instead of guessing, kept calls short and focused, and delivered a clear, structured plan without fluff or generic AI-style language.
Overall, I’m very happy with the outcome and would really recommend working with Muhammad.
KB
Kevin B.
May 28, 2026
Media Pipe. POC . CV Lacrosse
CM
Colton M.
May 18, 2026
Automated Drill Video Processing MVP - Phase 1
KS
Kristof S.
May 18, 2026
30 minute consultation
Good consultation
CM
Colton M.
Jan 30, 2026
30 minute consultation
About Muhammad
Computer Vision & AI Agent | Python - MediaPipe - YOLO Real-Time
100%
Job Success
Karachi, Pakistan - 12:23 pm local time
I've shipped 54+ that didn't.
⚙️YOLO Detection | Pose Estimation | Object Tracking | AI Agents | LLM Integration
Sports & Fitness AI | CCTV & Surveillance AI | Retail AI | Healthcare AI
You have a working concept... or a clear problem involving cameras, video, or image data. The challenge is making it fast, accurate, and stable under real-world conditions. Wrong framework choices. Inference too slow for live video. Models that break the moment lighting, angle, or environment changes. And systems that detect things but can't reason about them or act on them autonomously. That's exactly where most builds stall.
I design and build real-time computer vision pipelines that go all the way... from model training to live deployment... and increasingly, from visual perception to autonomous AI agents that understand, decide, and narrate.
LLM APIs (OpenAI, GPT-4o, Gemini, Claude) | AWS (EC2, S3, Lambda) | Azure Cloud Services | MLOps & API Integration | Model Deployment & Scaling
While most CV engineers stop at training the model, I go further:
→ High-speed inference optimization using TensorRT, ONNX, OpenVINO, FP16/INT8 (up to 5× faster)
→ LLM agents integrated with vision pipelines for alerts, reasoning, and automation
→ Mobile AI deployment using Core ML (iOS) and TFLite (Android) with 10+ shipped apps
→ Edge AI deployment on Jetson, OpenVINO, CUDA, and embedded systems
→ End-to-end pipelines: data → training → optimization → real-time deployment
Key Accomplishments:
⭐ $5M+ revenue from AI solutions
⭐ 100+ computer vision systems delivered
⭐ Built and launched 2 SaaS products
⭐ Real-time sports AI (7+ sports, 15+ teams)
⭐ 10+ mobile AI apps (iOS Core ML, Android TFLite)
⭐ Production AI for surveillance, industrial & safety use cases
⭐ Medical imaging AI deployed in 5+ hospitals
⭐ Up to 5× faster inference (ONNX, TensorRT, FP16/INT8)
⭐ Large-scale tracking & re-ID (1M+ labeled data)
⭐ Agentic AI systems for autonomous decision-making
If you have read this far, please note that I appreciate you taking the time to learn about me. Personally, it’s been an amazing journey and knowledge exercise to get to this level of competence in AI and software development.
Domain Expertise:
✅ athlete tracking | shot detection | scoring | drill analysis | pose estimation
✅ defect inspection | PPE compliance | staff monitoring | meter reading | quality control
✅ ANPR | crowd monitoring | people counting | intrusion detection | perimeter security
✅ tumor detection | ultrasound | X-ray/CT analysis | lesion segmentation | medical imaging
✅ aerial monitoring | traffic flow | license plate recognition | vehicle & accident detection
✅ customer analytics | receipt extraction | shelf monitoring | inventory tracking
Tech Stack:
YOLOv5–YOLOv8–YOLOv11, Detectron2, MMDetection, DeepSORT, StrongSORT, MediaPipe, OpenPose, Pose Estimation, Action Recognition, Segmentation (semantic & instance), OCR, anomaly detection, object tracking, PyTorch, TensorFlow, TFLite, Core ML, OpenCV, FastAPI, Flask, ONNX, TensorRT, OpenVINO, CUDA, AWS, Azure, GCP, edge AI, mobile AI, real-time inference, video analytics, AI automation, LLM integration (GPT-4o, Claude, Gemini, Groq), LangChain, LangGraph, CrewAI, RAG systems.
💬 If your project involves cameras, video, or images... and you need it fast, accurate, fully deployed, and intelligent enough to reason and act autonomously... I am the engineer you are looking for.
Steps for completing your project
After purchasing the project, send requirements so Muhammad can start the project.
Delivery time starts when Muhammad receives requirements from you.
Muhammad works on your project following the steps below.
Revisions may occur after the delivery date.
Requirement Collection
Client purchases the project and shares requirements (model files, target platform, export format, performance goals, test data).