Languages
01 / 8 technologies- Python
- TypeScript
- JavaScript
- Go
- Java
- C++
- SQL
- Rust
05 / Technology
A modern stack behind every build
A transparent view of the languages, frameworks, and platforms we use to build and run reliable production systems.
Production topologyCLIENT→API GATEWAY→AGENT ORCHESTRATOR→RAG · VECTOR DB→LLM / SLM→ON-PREM GPU
The stack, layer by layer
Eleven layers, one spatial system
Scroll and the instrument turns through all eleven layers. The active panel reads out the tools in that layer; below it, every named technology gets a textured 3D marker in the complete visual index.
Full 3D technology index
AI / ML
02 / 7 technologies- PyTorch
- TensorFlow
- scikit-learn
- Hugging Face
- XGBoost
- OpenCV
- ONNX
LLM & Agents
03 / 8 technologies- LangChain
- LangGraph
- LlamaIndex
- CrewAI
- AutoGen
- MCP
- vLLM
- Ollama
Vector & Data
04 / 7 technologies- Pinecone
- Weaviate
- Milvus
- pgvector
- FAISS
- Spark
- Airflow
Frontend & 3D
05 / 7 technologies- React
- Next.js
- Three.js
- React Three Fiber
- WebGL
- GSAP
- Tailwind
Backend
06 / 6 technologies- Node.js
- NestJS
- FastAPI
- Django
- GraphQL
- gRPC
Mobile
07 / 3 technologies- React Native
- Flutter
- Progressive Web Apps
Databases
08 / 5 technologies- PostgreSQL
- MySQL
- MongoDB
- Redis
- Elasticsearch
Cloud & DevOps
09 / 8 technologies- AWS
- Azure
- GCP
- Docker
- Kubernetes
- Terraform
- GitHub Actions
- CI/CD
MLOps
10 / 5 technologies- MLflow
- Kubeflow
- Model registry
- Monitoring
- Drift detection
Security
11 / 5 technologies- OAuth / SSO
- RBAC
- Encryption
- OWASP practices
- Secrets management
FAQ
Questions, answered
What technology stack does Kenzed Tech Lab use?
Eleven layers covering languages, AI/ML, LLM and agent frameworks, vector and data stores, backend, frontend and 3D, mobile, databases, cloud and DevOps, MLOps, and security. Python, TypeScript, Next.js, PostgreSQL, LangChain, LangGraph, Docker and Kubernetes are among the tools in regular use.
Which vector databases does Kenzed Tech Lab work with?
Pinecone, Weaviate, Milvus, pgvector and FAISS, with hybrid and semantic search. The choice depends on data volume, latency targets and whether the index has to stay inside the client's own infrastructure.
Is retrieval-augmented generation (RAG) part of what Kenzed builds?
Yes. RAG grounded in a private knowledge base is a standard part of the agent work, alongside tool use, function calling and Model Context Protocol (MCP) integrations.
Can Kenzed Tech Lab deploy models on-premise?
Yes. Kenzed Tech Lab operates in-house AI compute and builds on privately hosted and self-hosted models, which is the route taken when data cannot be sent to a third-party API.