Senior AI Engineer - RAG & Retrieval Engineering - Hungary

Senior AI Engineer – RAG & Retrieval Engineering expert needed

Location: Budapest, Hungary – Hybrid

Working model: Regular on-site presence in Budapest required

Focus: Production RAG | Retrieval Engineering | LLM Systems | AI Product Engineering

Note…..his is not a “plug an LLM into an API” role.

We’re looking for an experienced AI Engineer with genuine, hands-on retrieval engineering expertise to join a product-led technology business building intelligent AI capabilities into an established enterprise software platform.

The core challenge is complex: taking a large and continuously growing pool of real-world enterprise data and ensuring an AI system can retrieve the right information, for the right context, with measurable accuracy.

We therefore need someone who has genuinely built, debugged and improved production RAG systems from first principles – not simply connected documents to an off-the-shelf framework or LLM API.

You’ll take significant ownership of the retrieval layer and continuously improve its accuracy, reliability and ability to generate genuinely grounded recommendations.

Your primary focus will be retrieval engineering, including:

• Designing and building production-grade RAG systems from scratch
• Developing ingestion, chunking, embedding, indexing and retrieval strategies
• Implementing hybrid retrieval, combining semantic/vector search, metadata filtering, keyword search and reranking
• Building multi-stage retrieval pipelines designed around accuracy rather than simply semantic similarity
• Selecting and evaluating embedding models and vector-search approaches
• Working with vector databases at production scale
• Extracting clean, structured metadata from messy enterprise data to improve retrieval accuracy
• Building evaluation harnesses that measure retrieval precision against ground-truth datasets
• Analysing retrieval failures and systematically improving performance
• Developing validation logic that determines whether retrieved information is genuinely relevant to the user’s context rather than merely appearing plausible
• Building human-review and feedback workflows that continuously improve retrieval and validation quality

Wider AI engineering:

You’ll also contribute to the wider AI platform, including:

• Prompt engineering and structured prompt architecture
• Context engineering
• AI-powered chat and recommendation workflows
• Production LLM integration
• Agentic workflows and tool calling
• AI evaluation and observability
• Shared AI infrastructure used across multiple product teams

You’ll work closely with experienced software engineers, architects and product teams, so we’re looking for someone who can contribute technically while also thinking critically about how AI systems should behave in production.

What we’re specifically looking for:

We’re deliberately less interested in how many years you’ve worked in AI and much more interested in what you’ve actually built.

You should be able to talk in depth about decisions you’ve personally made around:

• Chunking strategy – how and why you structured source information
• Embedding model selection
• Vector indexing and retrieval
• Semantic vs keyword retrieval
• Hybrid search
• Reranking
• Retrieval precision and relevance
• Ground-truth datasets
• RAG evaluation
• Hallucination and grounding
• Context engineering
• Debugging retrieval failures
• Improving production AI systems over time

If your RAG experience is primarily LangChain + vector database + LLM API, without deeper ownership of the retrieval architecture, this probably isn’t the right opportunity.

Essential technical experience:

• Strong Python/software engineering background
• Commercial experience building production LLM-powered products or systems
• Deep hands-on RAG/retrieval engineering experience
• Vector databases and semantic search
• Hybrid and multi-stage retrieval
• Reranking techniques
• Retrieval evaluation against ground truth
• Prompt and context engineering
• REST APIs
• CI/CD and production deployment

Experience with any of the following would be particularly valuable:

• Graph databases / GraphRAG
• LangGraph or similar orchestration frameworks
• Multi-agent systems
• Tool calling
• MCP
• State management
• AI observability/evaluation platforms
• Shared AI infrastructure or AI platforms

We’re particularly interested in engineers who have worked within product-led technology environments, where you’ve owned AI capabilities beyond the initial prototype and seen how they behave with real users and real production data.

You might currently be working for an AI product company, SaaS business or product engineering organisation, or you may have come from consulting and now want the opportunity to own one AI product deeply rather than moving between short-term customer projects.

Either way, we’re looking for someone who thinks like a product engineer: build it, measure it, understand why it fails, improve it and own the outcome.

Why consider it?:

You’ll be joining a dedicated AI function where retrieval isn’t a side feature – it’s central to the intelligence of the product.

Rather than inheriting a finished RAG implementation, you’ll have genuine influence over how the retrieval architecture is designed, evaluated and evolved.

You’ll be solving difficult AI engineering problems against complex real-world enterprise data, with the opportunity to see the impact of your work compound as the product and its knowledge base grow.

Budapest – Hybrid

You must be based within realistic commuting distance of Budapest and comfortable working regularly with the team on-site.

If you’ve genuinely built sophisticated RAG/retrieval systems in production and can explain why your retrieval architecture works, how you measure it, and what you do when it doesn’t, I’d be very interested in speaking with you.