Unclaimed profileBuilt from public GitHub and Hugging Face activity. Vitali has not joined Lounge yet.Claim it
Not on Lounge yet
Vitali Ryumshin
developerlet's AI do DevOps job
Public projects
evidra
Open-source MCP execution-evidence recorder: wraps one upstream MCP server and records what an agent declared, what the proxy observed, and what it reported — a signed local chain for humans to reconcile.
Goaiai-agentaudit
evidra-lock
MCP Kill-switch for AI agents. Validates infrastructure operations before execution. Fail-closed. Evidence-backed.
Goaiai-agentai-guard
mlflow-enterprise-gateway
MLflow Multi-Tenancy & Governance Gateway adds tenant isolation, access control and audit logging to self-hosted MLflow without modifying MLflow itself. Designed for Kubernetes, private cloud and regulated environments where SaaS ML platforms are not an option.
Pythonaudit-loggingiammachine-learning
dsh-model-pricing
Model pricing & capability board for DeepSeek Harness: ~7,250 models across 213 providers, honest lowest-listed comparison, session costs priced by the route you actually called, and cache-leak attribution. Local-first, no server.
JavaScriptai-agentscodingdeepseek-harness
dsh-jev-subagent-dispatch
Cut LLM costs: route routine tasks to cheap subagent models (DeepSeek Harness plugin)
JavaScriptdeepseek-harnessdshdsh-plugin
evidra-bench
Evidra Bench runs live infrastructure scenarios for AI agents, MCP servers, and AI SRE tools, then shows pass/fail, unsafe passes, failure modes, artifacts, and reproducible evidence
Goagent-benchmarkai-agentsai-evaluation
Likely stack
MCP
