The problem ShifaAi solves
Problem it Solves
ShifaAI is an AI-powered molecule screening platform designed to tackle the challenges of traditional drug discovery, which is often slow, costly, and inaccessible to labs in underserved regions. By leveraging machine learning, ShifaAI rapidly analyzes and screens potential drug molecules, reducing dependency on expensive infrastructure while making the process faster and more affordable. The prototype, currently trained on 30,000 synthetic datasets for common cold treatments, demonstrates how AI can unlock new therapeutic possibilities. With support from pharmacists and researchers, ShifaAI has the potential to scale into a reliable healthcare assistant that accelerates drug discovery, makes treatments more accessible, and ensures safer, more efficient healthcare solutions.
Challenges we ran into
Challenges We Ran Into
- Synthetic Dataset Generation
Challenge: Since real medical datasets are hard to access due to privacy and regulation, we struggled with creating a reliable dataset.
Solution: We generated 30,000 synthetic samples through research and verified patterns to train our model.
2.Model Accuracy & Overfitting
Challenge: Our early machine learning models gave inconsistent predictions and tended to overfit.
Solution: We optimized the Decision Tree model by tuning hyperparameters and validating with cross-testing.
3.Resource & Infrastructure Limitations
Challenge: Running large-scale AI experiments is computationally expensive.
Solution: We used cloud resources and optimized our code to run efficiently on limited hardware.