Blog
Technical insights on building AI products, self-hosted infrastructure, and production ML systems.
Engineering8 min readFeatured
How We Built a Visual Product Search Engine with CLIP
A deep dive into building Trooply Search — a GPU-accelerated visual search engine using OpenAI's CLIP model, Qdrant vector database, and FastAPI. From architecture decisions to production deployment.
T
Trooply EngineeringCLIPvisual searchQdrantFastAPI+2
Infrastructure6 min readFeatured
Self-Hosted AI vs Cloud APIs: Why We Chose to Build on Our Own GPU
A practical comparison of self-hosted AI inference vs cloud APIs like OpenAI, Google Cloud AI, and AWS Bedrock. Cost analysis, latency benchmarks, and privacy implications for production workloads.
T
Trooply Engineeringself-hostedGPUcloud vs on-premiseAI infrastructure+1
Infrastructure7 min read
Running 22 Docker Containers on a Single Server: Our Infrastructure
How Trooply runs 6 AI products, 5 databases, and 11 infrastructure services on one machine. Docker Compose architecture, network isolation, resource limits, and lessons learned.
T
Trooply EngineeringDockerinfrastructureDevOpsself-hosted+1