Senior AI Engineer

**Role Overview** We are looking for a **Senior AI Engineer** to lead the development of enterprise-grade, autonomous agentic systems. In this role, you will build stateful AI workflows, standardize external tool connectivity using the **Model Context Protocol (MCP)**, and orchestrate complex business logic using frameworks like **LangGraph** combined with workflow platforms like **n8n**. The ideal candidate combines strong Python backend fundamentals with deep AI expertise. You are someone who already uses next-generation AI-assisted developer tooling (such as **Claude Code**) to ship clean, production-ready code with high velocity. **Key Responsibilities** - **Agentic Orchestration & Graph Workflows:** Design, build, and deploy multi-agent systems and stateful, cyclic reasoning workflows using **LangGraph** and **LangChain** (handling memory, subgraphs, recursion limits, and human-in-the-loop interrupts). - **Tool Layer Standardization (MCP):** Architect, deploy, and maintain custom **Model Context Protocol (MCP)** servers and client interfaces to give agents safe, standardized access to internal databases, APIs, and enterprise systems. - **Hybrid Automation Architecture:** Bridge custom Python microservices with workflow automation engines like **n8n** to automate complex end-to-end multi-app processes and webhook-driven event pipelines. - **Production API & Systems Development:** Develop robust, asynchronous microservices and endpoints (**FastAPI**) incorporating rate limiting, streaming responses (SSE/WebSocket), retries, and token-cost tracking. - **AI-Native Engineering:** Champion AI-assisted development practices by heavily utilizing **Claude Code** and other agentic developer CLI tools to rapidly scaffold, test, and refactor code. - **Agent Reliability & Guardrails:** Implement safety guardrails, schema validation, structured output enforcement, and automated agent evaluation suites (evaluating tool-call accuracy, task completion, and hallucination rates). **Required Skills & Qualifications** - **Core Engineering Experience:** 4+ years of professional backend software engineering in **Python**, with strong proficiency in asynchronous programming (asyncio, httpx), OOP, and clean architecture. - **Agentic Frameworks:** 2+ years of hands-on experience building autonomous agents and stateful graphs using **LangGraph** (managing checkpointers, state schemas, and branch conditions) and LangChain. - **Model Context Protocol (MCP):** Hands-on experience building, debugging, and deploying custom **MCP servers** (Stdio/SSE transports) and connecting them to agent runtimes. - **Modern Agentic Tooling:** Daily fluency with **Claude Code** or comparable agentic CLI tools for automated testing, debugging, and rapid development. - **Structured Outputs & Tool Calling:** Deep understanding of function/tool calling mechanics and schema enforcement libraries (e.g., **Pydantic**, Instructor, JSON Schema). - **RAG & Search:** Solid foundation in Retrieval-Augmented Generation (RAG) architectures: hybrid search (dense embeddings + sparse keyword search), re-ranking, and production vector stores (e.g., Qdrant, Weaviate, Pinecone, or pgvector). - **System Design & Resilience:** Strong system design fundamentals: API authentication, asynchronous task queues (Celery/Redis), rate-limiting, and error-recovery patterns for non-deterministic systems. **Preferred / Nice-to-Have** - **In-House LLM Serving:** Experience integrating with or deploying local/open-source models using engines like **vLLM, Ollama, or Triton Inference Server** on internal GPU infrastructure. - **Observability & LLMOps:** Experience setting up end-to-end agent tracing, cost analysis, and evaluation frameworks using **LangSmith, Arize Phoenix, or OpenTelemetry**. - **Cloud Platforms:** Experience running and containerizing AI services on **AWS, Azure, or GCP** using Docker and Kubernetes. - **MCP Open Source Ecosystem:** Active contributor to or early adopter of the open-source MCP community and servers. - **Frontend Basics:** Working familiarity with frontend frameworks (React, Next.js, or Tailwind) to quickly prototype internal agent-interaction UIs. *This is a hybrid work model, involves working from office and attending in person interviews.* **Why Join PalTech?** High Ownership & Visibility Modern Tech Stack Career Acceleration Learning-Driven Culture Collaborative Environment Hybrid Work Model If you're looking to build, scale, and lead systems that truly matter, PalTech provides the right ecosystem.