AI Thesis: Skin Disease Diagnosis
A+ grade and a job offer from a global firm during the defense
- Client
- Khalid AlMalki (student)
- Year
- 2025
- Duration
- 6 weeks
- Team
- 2 engineers
SkinNet AI
EfficientNet-B3 · v2.1
Input
Drop an image here
PNG, JPG · max 10MB
—
Diagnosis
Waiting for image…
Helped Khalid build a complete graduation project using CNNs to detect skin diseases, with academic documentation and a polished defense presentation.
The challenge
Khalid had a topic but didn't know where to start. The deadline was tight (6 weeks), and he wanted something that would impress the committee far beyond expectations.
The result
Khalid got A+ from the committee and a job offer from a global pharma company that observed the defense. He's now in their AI Research team in Dubai.
Our approach
- 01
Selected suitable dataset (HAM10000) + augmentation for 50K images
- 02
Tested 5 CNN architectures, picked EfficientNet-B3
- 03
Fine-tuning + ensemble learning to reach 94.2% accuracy
- 04
Built a Flask web app for interactive demo
- 05
Wrote 5 academic chapters (150 pages) + IEEE references
- 06
Defense prep with 30 likely questions
“Helped me with my AI graduation project on skin disease diagnosis. Got an A+ and a job offer during the defense panel.”