LLM Engineer

Hello Do you believe you found this opportunity by chance?At Nimber, we dont believe in chance we believe in curiosity, creativity, and people who build with purpose. If this role sparked your interest, great. Keep reading Were not just hiring were building a team of AI pioneers ready to shape the future of intelligent agents. If youre passionate about LLMs, agent architecture, and solving complex problems with cutting-edge tools, this is your moment. Position: LLM Engineer - AI Agent Development Location: Maia, Porto (On-site)Experience Level: Junior to Mid (1-4 years)Employment Type: Full-time About the Role Youll join our AI development team to refine existing agents and help build the next generation using MCP (Model Context Protocol). From hallucination control to multi-agent orchestration, youll be working at the forefront of conversational AI and customer support automation. Key Projects Current: Backoffice Agent Refinement Enhance agent capabilities and performance Optimize workflows using LangGraph for multi-step reasoning Implement hallucination detection and mitigation strategies Improve reliability, error handling, and output validation Upcoming: Frontend Agent & MCP Server Design real-time customer-facing agents Implement MCP server architecture for seamless model integration Build scalable infrastructure for concurrent interactions Develop robust conversation management and context preservation Key Responsibilities AI Agent Development & Architecture Design complex workflows with LangGraph Build LangChain pipelines for document retrieval Implement hallucination control and validation mechanisms Create scalable agent architectures for high concurrency Develop RAG systems with vector databases Advanced Agent Capabilities Orchestrate multi-agent systems Implement tool calling and function execution frameworks Design memory systems for long-term context Build agent evaluation and testing frameworks Optimize prompts and dynamic generation strategies Production & Scalability Apply horizontal scaling strategies Develop caching for embeddings and responses Implement load balancing and token cost optimization Build rate limiting and abuse prevention mechanisms Quality & Safety Create output validation pipelines Implement safety filters and content moderation Develop confidence scoring and uncertainty quantification Design human-in-the-loop workflows Build audit trails and logging systems Required Technical Skills AI/ML Frameworks & Tools LangChain & LangGraph for agent workflows Vector databases (PostgreSQL + pgvector) LLM APIs (OpenAI GPT-4, Claude, etc.) Embeddings for semantic search and retrieval Hallucination Control & Validation Fact-checking and ground truth verification Confidence scoring and output validation Guardrails and content filtering Scalability & Performance Async programming with Python asyncio Queue systems (Redis, Celery, RQ) Caching strategies and database optimization Load testing for high-concurrency scenarios Backend Development Python (2+ years), FastAPI or Flask PostgreSQL, vector DBs, ORM usage Authentication (JWT, OAuth) Docker and container orchestration Soft Skills & Attributes Creative problem-solving mindset Precision in prompt engineering and validation Empathy for user experience in conversational interfaces Analytical thinking and data-driven decision-making Strong communication in English and Portuguese Adaptability to fast-evolving AI technologies Commitment to quality and production-readiness What We Offer Competitive salary based on experience Budget for AI/ML courses and certifications Access to premium LLM APIs and dev tools Modern office in Maia with high-performance hardware Flexible working hours within core business times Opportunity to work on cutting-edge AI agent tech Coffee and snacks always available Work Environment This is an on-site position in our Maia office. We foster a collaborative space where AI engineers experiment, share discoveries, and iterate quickly. Our team thrives on continuous learning and pushing the boundaries of agent development. Technical Stack AI/ML: LangChain, LangGraph, GPT-4, Claude, Hugging Face Languages: Python (primary), TypeScript/JavaScript (frontend) Databases: PostgreSQL + pgvector, Redis Infrastructure: Docker, Kubernetes, AWS/GCP/Azure Monitoring: OpenTelemetry, Prometheus Development: FastAPI, Pydantic, pytest, black, mypy Technical Interview Topics Agent Architecture & Development LangGraph workflow design LangChain pipeline optimization RAG architecture and vector DB selection Multi-agent coordination patterns Quality & Safety Hallucination detection and mitigation Output validation and fact-checking Confidence scoring and safety filters Scalability & Performance Horizontal scaling and caching Async programming for concurrency Token optimization and cost control Integration & Production MCP protocol implementation Real-time conversation systems Error handling and observability Sample Technical Challenges You may be asked to: Design a LangGraph workflow for a multi-step support scenario Implement hallucination detection for agent responses Architect a system for 1000+ concurrent conversations Debug inconsistent agent outputs Optimize a RAG system for speed and accuracy Ready to stop being just another number?Apply now and become a Nimber where your ideas shape the future of AI.