Description:
The Role:
You will design intelligent, autonomous agents and own the full stack: from the cognitive architecture of multi-agent systems to high-performance backends and frontends that make agent reasoning transparent in real time.
AI & Agent Orchestration:
- Design and implement multi-agent workflows using frameworks such as LangGraph, CrewAI, or Semantic Kernel.
- Embed advanced reasoning patterns (ReAct, Plan-and-Execute, Reflexion) to reduce hallucinations and improve decision-making.
- Architect long-term memory and Retrieval-Augmented Generation (RAG) pipelines with vector databases (Pinecone, Milvus, Weaviate) to ground agents in real-world data.
- Integrate external tools securely through the Model Context Protocol (MCP), connecting agents to databases, APIs, collaboration platforms, and internal systems.
Backend Engineering:
- Develop high-concurrency, non-blocking backends with Python (FastAPI) or Node.js to handle long-running agent tasks without blocking the UI.
- Implement AgentOps and observability (LangSmith, Arize Phoenix) to trace, debug, and explain agent decisions.
- Build safety guardrails and deterministic fallback layers to intercept and correct unintended agent actions before production.
Frontend & User Experience:
- Create streaming interfaces (React, Next.js) that push real-time "thinking" states, tool-use logs, and incremental results to users.
- Design human-in-the-loop (HITL) approval components for human confirmation on high-stakes actions.
- Build collaborative dashboards that visualize multi-agent handoffs (for example, a "Researcher" agent passing data to a "Writer" agent).
What We're Looking For:
- Expert-level Python (AI/backend) and TypeScript (frontend and API safety).
- Deep experience with agentic frameworks (LangGraph, LangChain), frontend frameworks (React, Next.js), and backend frameworks (FastAPI).
- Proficiency in relational databases (PostgreSQL) and vector stores (Weaviate, Chroma, Pinecone).
- Hands-on experience with Docker, Kubernetes, and AI deployment platforms (Azure AI Foundry, AWS Bedrock, Vercel AI SDK).
- Proficiency in English for effective asynchronous collaboration in a remote, multicultural team.
Bonus Points:
- Contributions to open-source agentic or LLM projects.
- Experience with small language models (SLMs) for low-latency, on-device agent tasks.
Why You'll Love Working With Us:
- Fully remote with stable, full-time income in US dollars or local equivalent.
- Work at the cutting edge of agentic AI, 3D web, and Web3 with real ownership of architecture.
- Continuous learning culture with shared prompts, frameworks, and research.
- Inclusive, ego-free team where ideas matter.
What we're looking for
- Agentic AI: LangGraph, LangChain, ReAct, Plan-and-Execute, Reflexion, RAG, MCP, Pinecone, Milvus, Weaviate, Chroma
- Backend: Python, FastAPI, Node.js
- Observability & AgentOps: LangSmith, Arize Phoenix
- Frontend: React, Next.js, TypeScript, SSE/WebSockets
- Databases: PostgreSQL; vector stores such as Weaviate, Chroma, Pinecone
- DevOps & Deployment: Docker, Kubernetes, AWS Bedrock, Azure AI Foundry, Vercel AI SDK
- Other: AI deployment, AgentOps, observability, safety guardrails, strong written and verbal English for asynchronous remote work
