Qwen 2.5 7B on Ollama. Tool-capable, JSON-tuned chat model. A solid default for agent UIs (OpenCode, OpenClaw, OpenFang) that require tool calling.
Every EasyEnv recipe spins up in seconds on a real Linux VM-not a stripped-down sandbox. The Ollama Qwen 2.5 7B recipe is provisioned by an open Ansible role, so the machine that boots for you is reproducible, inspectable, and matches what you would get in production.
$ easyenv workspace create --recipe ollama_qwen2_5 --name ollama_qwen2_5-demo
Provisioning Ollama Qwen 2.5 7B...
Workspace ready in ~45s
$ easyenv workspace ssh ollama_qwen2_5-demo
Connected. You're on the machine.Stop guessing from resumes. Drop candidates into a real Ollama Qwen 2.5 7B workspace, watch them debug, deploy and operate it, score the result automatically, and replay the session. We also evaluate how they work with AI.
Builds and deploys AI applications, integrates LLMs into products, and manages self-hosted AI infrastructure
PostgreSQLDevelops and deploys machine learning models and AI systems for production environments
Self-hosted LLM inference with Ollama and OpenWebUI for private AI workloads
Self-hosted AI document assistant powered by AnythingLLM and Ollama
Automated AI workflows combining LLMs with n8n orchestration and PostgreSQL
Machine learning and data science environment
