All projects
NutriTrov AI — Nutrition Discovery
AI-powered nutrition & healthy food discovery platform. A conversational assistant explains WHY each food is recommended. Semantic search via pgvector finds intent-aware results, with nearby healthy restaurants on a map and full nutrition breakdown with AI health scores.
Angular 19ASP.NET CoreFastAPILangChainpgvectorPostgreSQLRedisDockerAzure
What it does
Key features
- Conversational AI nutrition assistant — always explains WHY a food is recommended
- Semantic search with pgvector — "crunchy gym recovery snacks" finds relevant products intelligently
- Nearby healthy food discovery — restaurants, cafes, stores on an interactive map
- Full nutrition intelligence per product — macros, vitamins, allergens, AI health score
- Personalised profiles — goals, allergies, dietary preferences, calorie targets
- Explainable recommendations — LangChain reasoning surfaces the exact matching criteria
- Admin dashboard — CRUD for products, nutrition facts, stores, and analytics
Under the hood
Architecture
FrontendAngular 19 + Signals + Angular Material + Tailwind
API GatewayASP.NET Core Clean Architecture — CQRS + MediatR, JWT, EF Core
AI ServiceFastAPI + LangChain + OpenAI + SentenceTransformers — embeddings, reasoning
DatabasePostgreSQL + pgvector — products, nutrition_facts, embeddings, conversations
CacheRedis — recommendation cache, session management
DeploymentDocker; Azure App Service + Azure PostgreSQL; Vercel frontend
Interested in work like this?
Get in touch