Building AI agents and LLM infrastructure. I work across the full stack — TypeScript/Bun on the agent runtime side, Python on the framework and tooling side.
Currently focused on agentic workflows, multi-provider LLM orchestration, and developer tooling for AI-native applications.
jarvis — Personal AI assistant
TypeScript · Bun · ReAct loop · Multi-provider LLM · MCP · Channel integrations
A persistent, tool-augmented CLI agent with session memory, Dream consolidation, and an OpenAI-compatible API server. Inspired by nanobot.
bun install -g @kindred/jarvis
jarvis agenttinyflow — LLM agent framework
Python · Pydantic · Asyncio · ChromaDB · Qdrant
A lightweight, provider-agnostic Python framework for building LLM agents. Designed around clean primitives: a Thinking→Acting→Observing loop, a @tool decorator, and structured JSON streaming.
uv add tinyflow-llm| Layer | Stack |
|---|---|
| Agent runtime | TypeScript, Bun, Node.js |
| Agent framework | Python 3.12+, Pydantic, Asyncio |
| LLMs | DeepSeek, OpenAI, Anthropic, Gemini, OpenRouter |
| Vector stores | ChromaDB, Qdrant |
| Embeddings | OpenAI, sentence-transformers |
| Integrations | Feishu, Discord, Telegram, Slack, Email |
| Infra | MCP servers, OpenAI-compatible API |
- GitHub: @kindredzhang
- Interested in AI agents, developer tooling, or open source collaboration — feel free to open an issue or reach out.
