Browse categories
Explore
Fiverr Pro
English
$
USD
I will build custom computer vision and object detection systems
Know What to Build Before Investing in AI, Data, Software, and Cloud Systems
Hello, I’m Drew, the founder of Lykios, a technology company based in San Diego, California. We engineer intelligent, secure, and scalable digital systems for clients worldwide. Our services include A...
Level 1
Has met certain performance criteria and shows strong potential in the marketplace.
About this Gig
Teach a camera to see what matters to your business: defects, products, vehicles, animals, and more.
LYKIOS builds custom computer vision systems using your images and footage. We prepare the data, train and test a model, then deliver a working solution you can use.
What we build:
- Object detection, segmentation, and image classification
- Visual quality inspection and defect detection
- Video analytics: tracking, counting, zones, and alerts
- Inference APIs and camera pipelines
We work with tools including PyTorch, YOLO, OpenCV, and FastAPI. You receive the agreed model, source code, performance report, and documentation. We measure results on your data and explain where the model works well and where it needs more examples.
Choose Basic for a proof of concept, Standard for a custom model with one video feature and an API, or Premium for a deployed system covering up to two cameras. Scope, accuracy, and speed depend on your footage and hardware.
Message me with sample images or a short video, what you need to detect, and where the system will run. Ill recommend the right package or a custom quote.
APIs:
Google Cloud Vision API
•
Other
Programming language:
Python
•
R
•
SQL
•
Java
Tools:
Jupyter Notebook
•
OpenCV
•
MLflow
•
CVAT
•
PyTorch
Frameworks:
Scikit-learn
•
DeepPy
•
Other
My Portfolio
Other Data Science & ML Services I Offer
FAQ
What do you need from me to get started?
Send sample images or footage, what you want detected, and where the system will run. If you have labeled data, tell me how many images you have. Please message me before ordering so I can confirm the right package.
Can you work with a small image dataset?
Yes. We can test a pretrained model or fine-tune one on your images. Results depend on the number and variety of examples, so we’ll assess feasibility before promising a performance target.
What is included in the Basic package?
A proof of concept for up to 3 object types using up to 300 buyer-provided images, including a trained model, demo script, results report, and source code. It does not include a live camera pipeline or deployment.
Can the system analyze a live camera feed?
Yes. Standard includes one agreed video feature, such as tracking or counting. Premium covers a deployed pipeline for up to 2 cameras. Send your camera type and sample footage so we can confirm compatibility.
Will it work at night or in poor lighting?
We can train and test with low-light or infrared footage if you provide examples from those conditions. We’ll report the results separately; daytime performance does not guarantee night performance.
Do I receive the source code and model?
Yes. You receive the source code and model files included in your package, along with documentation. Third-party tools and models remain subject to their applicable licenses.
