SWOC ‘25 | Code. Contribute. Collaborate.
Building AI-Powered Solutions: From Smarter Healthcare to Intelligent E-Commerce!
Created on 23rd February 2025
•
SWOC ‘25 | Code. Contribute. Collaborate.
Building AI-Powered Solutions: From Smarter Healthcare to Intelligent E-Commerce!
The problem SWOC ‘25 | Code. Contribute. Collaborate. solves
I contributed to impactful projects that make healthcare smarter, e-commerce effortless, and data more powerful!
1. TelMedSphere
Need a doctor? TelMedSphere makes healthcare hassle-free! Patients can connect with doctors via video consults, manage health records, and get prescriptions. It also features a disease prediction model that analyzes symptoms and provides possible solutions. Meanwhile, doctors can streamline appointments with ease.
2. SocioSell
What if your product pictures could speak for themselves? SocioSell transforms product images and videos from social media into ready-to-sell listings! Using AI, it analyzes media, extracts product details, and generates listings automatically—making online selling seamless and hassle-free.
3. Dataverse
Data is power—when it’s easy to use! Dataverse transforms raw data into interactive visuals, simplifies financial management, and leverages predictive analytics to provide actionable insights, all while ensuring secure storage.
Challenges I ran into
TelMedSphere
- Authentication & Frontend Integration: Integrating Firebase credentials and ensuring a smooth user experience was challenging. After troubleshooting API configurations and debugging authentication flows, I successfully established a seamless login system.
- Scaling Media Storage & Notifications: Handling large media files while maintaining Twilio WhatsApp notifications required overcoming API rate limits. By optimizing asynchronous processing and leveraging Cloudinary’s compression and caching, I achieved 98% reliability in media uploads and message delivery.
SocioSell
- Optimizing AI-Generated Listings & Data Validation: The Gemini AI API’s response structure was mostly unusable (~5% useful), making automated product listing impractical. By fine-tuning prompts, structuring AI responses, and implementing Pydantic-based validation, I improved product detection accuracy to ~60% and increased structured usability to ~75% for reliable listings.
- Efficient Query Optimization: Slow database queries affected performance. After testing multiple indexing strategies, I achieved a ~55% improvement, significantly enhancing query efficiency.
- Automating Data Insertion: Manually adding test data was inefficient. I developed a single-command script to insert all necessary sample data into the database, reducing manual effort by 98% and ensuring consistency.
Key Takeaways from SWOC
Beyond technical challenges, SWOC reinforced the importance of clear communication, aligning with project expectations, and structured collaboration. Moving from personal projects to open-source teamwork introduced challenges like resolving merge conflicts, refining PRs, and adapting to repository workflows. This experience significantly enhanced my technical adaptability, structured collaboration, and problem-solving skills.