EdTech platform feature development coupled with cost optimisations of ~38%

TC Global - a global education platform connecting students with universities faced two critical issues:

⚠️ Performance degradation as user base grew to 2.5M students

⚠️ 95% increase in cloud costs within just two months

The client needed comprehensive platform optimization with advanced search capabilities, improved data quality, and an ML recommendation system while simultaneously reducing infrastructure expenses.

✅ 44% Performance Enhancement: Successfully replaced legacy search functionality with a modern ElasticSearch implementation featuring fuzzy search capabilities and complex indexing systems enhanced by NLP.

✅ Architecture Modernization: Redesigned the application with asynchronous processing framework, containerization and horizontal scaling to meet strict performance requirements.

✅ Advanced Recommendation System: Implemented a neural network recommendation system for suggesting relevant courses and universities based on user profiles, inspired by triplet loss-based face detection architectures.

✅ Explainable AI: Introduced a game-theoretic approach for model explainability that maintained high accuracy while providing business stakeholders with result certainty metrics.

✅ 38% Infrastructure Optimization through:

➔ Implementation of AWS Graviton2 processors with 44% better price/performance ratio ➔ 50% reduction in Elasticsearch data nodes ➔ 75% reduction in ECS tasks during low-traffic periods through Auto Scaling ➔ Optimized database connection handling through improved batch processing

The solution maintained full application performance despite significant resource reduction and provided enhanced scalability during high-traffic scenarios.