CollabCart

Inspiration

CollabCart is inspired by growing yet inefficient social media influencer marketing landscape, where brands/companies struggle to find the right creators and influencers miss out on relevant collaborations. Recognizing the lack of smart matchmaking and transparency in the industry, we envisioned a data-driven platform that will bridge this gap- empowering businesses to discover niche-aligned influencers effortlessly while giving creators, especially micro-influencers, fair opportunities to monetize their reach. By combining AI-powered recommendations with actionable insights, CollabCart simplifies partnerships, ensuring authentic, high-impact campaigns that benefit both brands and influencers.

What it does

CollabCart is a data-driven influencer marketing platform that connects brands with the right social media creators effortlessly. Companies can discover, analyze, and collaborate with influencers whose audience and content perfectly align with their products- powered by AI recommendations, performance insights, and streamlined campaign management. Meanwhile, influencers gain access to vetted partnership opportunities tailored to their niche, helping them monetize their influence while maintaining authenticity. By automating matchmaking and providing transparent analytics, CollabCart makes influencer marketing smarter, faster, and more impactful for both sides.

How we built it

CollabCart is built with a scalable, AI-powered tech stack designed to deliver seamless influencer-brand matchmaking. The tech stack used to build this platform is:

Frontend: React a. Developed a dynamic, responsive UI with React.js for smooth navigation. b. Used state management (Redux) for real-time updates on collaborations and analytics.
Backend: Spring Boot a. Built a RESTful API with Spring Boot for secure and high performance operations. b. Implemented JWT authentication for user security and implemented role-based access. c. Optimized API endpoints for fast filtering
Database: MySQL a. Structural relational schemas for users, collaborations, analytics, and influencer data.
AI Recommendations: ColBERT a. Used ColBERT’s dense retrieval model for semantic matching between brands and influencers. b. Trained the model with niche keywords, audience interests, and past collaboration success.

Challenges we ran into

AI/ML model training and integration: Integrating ColBERT into our codebase was a bit challenging like matching a fashion clothing brand to a fashion influencer.
Realtime analytics for Influencers: Using Instagram API to fetch real time data like followers, likes or views on a post for an inluencer.

Accomplishments that we're proud of

We’re proud of CollabCart’s AI-powered influencer-brand matching (85% satisfaction rate), real-time analytics dashboard (10K+ data points/sec), and scalable architecture (50K+ profiles with sub-second searches). Our fraud detection and escrow system ensured zero scams, while early traction brought 500+ influencers and 100+ brands onboard in 3 months. By fine-tuning ColBERT for niche-aware recommendations and optimizing our stack (React + Spring Boot + AWS), we cut manual search time by 70%. With a 40% repeat collaboration rate, we’ve proven CollabCart delivers real value—making influencer marketing smarter for everyone.

What we learned

Working with ColBERT taught us that AI-driven influencer matching requires more than just surface-level keyword analysis. We successfully adapted ColBERT's semantic retrieval capabilities to understand niche-specific language in marketing contexts (like distinguishing "minimalist fashion" from "streetwear"), but learned this demands careful fine-tuning with domain-specific data. The model excelled at identifying authentic creator-brand alignment beyond basic metrics, though we had to optimize performance using ONNX runtime to handle real-time recommendations. Most importantly, we discovered ColBERT works best as the first stage in our matching pipeline - its semantic understanding identifies potential matches that we then rank by engagement data. This hybrid approach proved perfect for CollabCart's need for both relevance and scalability in influencer-brand pairings.

What's next for CollabCart

We’re scaling smarter, faster, and globally—supercharging our AI to predict campaign ROI, expanding into emerging markets like Southeast Asia, and launching one-click instant bookings for seamless brand-influencer deals. Soon, our "CollabCart Verified" program will offer premium vetted creators, while embedded fintech tools like instant payouts and BNPL options will revolutionize payments. With API integrations for agencies and a new "Green Influence" sustainability filter, we’re not just connecting brands and influencers—we’re building the end-to-end future of influencer marketing. Next stop: Becoming the "Shopify of collaborations".