Machine LearningFeatured
Travel Recommendation System
ML-powered travel recommendations
Machine LearningRecommendationsFastAPIStartup
Performance Metrics
Role
Technical co-founder
Team
2-3 people
Duration
2.5 years
Overview
Co-founded travel tech startup (Jan 2022 - Jul 2024):
**Role**: Technical co-founder, owned technical stack
**What We Built**:
- Full-stack application: React frontend, FastAPI microservices, PostgreSQL
- Hybrid recommendation engine: collaborative filtering (ALS) + LightGBM ranking + content-based
- Integrated external APIs: hotel and flight data
- Implemented Redis caching to reduce latency
- Ran A/B tests to measure engagement improvements
**Challenges**:
- Cold start problem for new users
- Data sparsity in early days
- Balancing recommendation quality vs. diversity
- Working with limited resources (2-3 person team)
**What I Learned**:
- Product development and user research
- Making engineering trade-offs with resource constraints
- Remote team coordination
- Not everything that works in theory works in practice
**Reality check**: User engagement improved but product-market fit was challenging. Project provided valuable learning experience.
Technologies Used
PythonFastAPIscikit-learnLightGBMPostgreSQLRedisReactAWS
Project Timeline
January 2022 - July 2024