Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

Randstad

NewBe an early applicant

Python AI Engineer

Posted Updated

Summary

Design, build, and operate production-grade AI applications and agentic workflows—RAG pipelines, LLM orchestration, MCP integrations—for a Dublin-based client, using Python, FastAPI, LangGraph, LangChain, MongoDB, and Airflow. Permanent, hybrid role (3 days/week on-site); no finance background required.

🚀 Python AI Engineer

  • 📍 Location: Dublin (Hybrid – 3 days/week on-site)
  • 💼 Employment Type: Permanent / Full-Time


💡 About the Role


We are seeking two experienced Python AI Engineers to design, build, deploy, and operate secure, production-ready AI applications and agentic workflows.

🚫 No Finance Experience Needed: We are looking for strong software developers who deeply understand working with Python and building clean, reliable backend systems—you do not need a background in finance or financial services to succeed here.

In this role, you will own the end-to-end engineering lifecycle of AI-enabled services—moving capabilities from early proof-of-concept into controlled, scalable production environments. You will combine large language models (LLMs), enterprise data, and deterministic software components using Python, FastAPI, LangGraph, LangChain, MCP, MongoDB, and Apache Airflow.


🛠️ Key Responsibilities


  • 🤖 Production AI & Agentic Workflows: Design and implement reliable, production-grade AI applications, copilots, and stateful agentic workflows aligned with measurable business outcomes.
  • ⚡ API & Service Engineering: Build maintainable Python services and expose AI capabilities through secure, versioned REST APIs using FastAPI.
  • 🔍 RAG Pipeline Architecture: Design and optimize end-to-end Retrieval-Augmented Generation (RAG) pipelines—covering document ingestion, chunking, metadata tagging, embeddings, hybrid retrieval, reranking, citations, and automated evaluation.
  • 🧠 LLM Orchestration & Governance: Integrate and invoke commercial or open-source LLMs via SDKs or model gateways, implementing structured outputs, tool/function calling, retries, rate limiting, token/cost controls, and model fallbacks.
  • 🔗 LangGraph & LangChain Implementation: Use LangGraph to architect stateful, controllable workflows (with checkpoints and human-in-the-loop capabilities) and apply LangChain pragmatically for model, tool, and retriever integrations.
  • 🔌 Model Context Protocol (MCP): Build and integrate MCP servers and clients to securely expose enterprise capabilities via MCP tools, resources, and prompts with strict schema validation.
  • ⏱️ Data Pipelines & Scheduling: Develop and operate Apache Airflow DAGs for scheduled data ingestion, embedding refreshes, batch evaluations, and data quality checks.
  • 🗄️ Database Design: Design MongoDB document models, indexes, and access patterns to manage application state, conversation history, metadata, and operational logs.
  • 🛡️ Production Operations & Security: Implement automated testing (pytest), CI/CD pipelines, container deployments, structured logging, tracing, monitoring, incident-ready runbooks, and security/privacy guardrails (such as prompt-injection mitigations and least-privilege access).
  • 🤝 Cross-Functional Collaboration & Documentation: Partner with product owners, data engineers, security architects, and SMEs to translate requirements into safe technical designs, documenting architecture decisions, model limitations, and data lineage.


🎯 Required Qualifications & Technical Skills


Core Background


  • 🐍 Python-First Software Developers: Strong software engineering background with deep, practical mastery of Python—no finance domain experience is required.
  • 📈 Production AI Delivery: Typically 4+ years of software, data, or ML engineering experience (including at least 1+ year delivering LLM-enabled applications), with a proven track record of taking AI capabilities from proof-of-concept (PoC) into controlled production deployments.
  • 🏗️ Engineering Judgment: Ability to structure modular, reusable, and maintainable services rather than notebook-only code, knowing when to use deterministic code versus LLMs or agents.


Essential Technical Skills


  • 🐍 Python Engineering: Strong proficiency in Python 3, including type hints, packaging, dependency management, asynchronous programming (asyncio), error handling, structured logging, and unit/integration testing (pytest).
  • 🌐 APIs & Web Frameworks: Hands-on experience building REST APIs with FastAPI and Pydantic (validation, dependency injection, async endpoints, OpenAPI, authentication/authorization) and familiarity with API testing tools such as ReadyAPI.
  • 🕸️ Agentic Frameworks: Practical experience modeling stateful AI workflows as graphs using LangGraph (nodes, edges, conditional routing, checkpoints, human-in-the-loop, failure recovery) and working knowledge of LangChain (or LlamaIndex) abstractions without over-engineering.
  • 🧩 Model Context Protocol (MCP): Hands-on understanding of MCP architecture, including building or integrating MCP servers and clients.
  • 📚 RAG & Information Retrieval: Practical delivery of RAG systems, including parsing, chunking strategies, vector/lexical/hybrid search, metadata filtering, reranking, prompt grounding, and measurable retrieval/answer evaluation.
  • 💬 LLM Invocation: Experience with commercial or open LLM SDKs/APIs, prompt engineering, structured outputs, tool/function calling, streaming, context-window management, and latency/cost optimization.
  • 💾 Databases (NoSQL & SQL): Strong understanding of MongoDB (collections, document modeling, schema validation, indexing, aggregation pipelines, transactions, query performance; Atlas experience is a plus) alongside solid SQL and data-processing fundamentals.


✨ Desirable / Good-to-Have Skills


  • 🌬️ Apache Airflow: Experience authoring reliable DAGs with clear task boundaries, retries, idempotency, backfills, secrets management, and operational recovery.
  • ☁️ Cloud & Model Gateways: Experience with Azure OpenAI, Anthropic, AWS Bedrock, RedHat OpenShift, or infrastructure-as-code.
  • 🔎 Advanced Vector & LLMOps Tooling: Familiarity with MongoDB Vector Search, Elasticsearch/OpenSearch, Milvus, or evaluation/tracing tools like Galileo
  • 🏛️ Platform Architecture: Familiarity with the Blueverse GenAI Architecture Platform or similar enterprise AI platforms.

Skills

Apply

See also

AI Engineering jobs by country — openings, pay and top skills →