Engineering
ML / AI Engineer (m/f/d)
Berlin
- On-Site Full-time Senior
- Mid
- Junior
- Entry Competitive salary + meaningful equity/upside
About The Role
neospaces is a new kind of commercial real estate brokerage. Proprietary technology. Personal dedication. Nationwide reach. For brands that want the perfect location — and owners looking for the perfect tenant. The AI layer is our moat. Matching tenants to spaces, qualifying leads, extracting structure from exposés and prose requirements, surfacing the right property at the right moment — this is where the product stops being a database and starts being an advantage. We're looking for an ML / AI Engineer who ships AI into production and is honest about what works: measured, evaluated, and improved — not demoed.
Responsibilities
- Build and own the matching engine: tenant requirements to spaces, ranked and explainable
- Design retrieval pipelines — embeddings, vector search, hybrid ranking
- Build structured extraction from unstructured sources: exposés, PDFs, listings, emails
- Design agentic workflows for qualification, enrichment, and research
- Build the evaluation layer: datasets, metrics, regression tests — so quality is measured, not guessed
- Get models into production and keep them there: latency, cost, monitoring, fallbacks
- Work directly with brokers to turn domain knowledge into system behavior
Requirements
- Hands-on experience putting LLMs or ML models into production — not just notebooks
- Solid Python and/or TypeScript, and comfort with the surrounding engineering work
- Experience with retrieval: embeddings, vector databases, ranking, hybrid search
- You evaluate rigorously and can tell a real improvement from a lucky prompt
- Pragmatic about tradeoffs: cost, latency, and accuracy are all constraints
- Bonus: agentic frameworks, Claude API, structured extraction at scale, recommender systems
- Bonus: experience with messy real-world data in a vertical domain
- Open to strong Entry and Junior candidates who can show real shipped work
- Business-fluent English; German a plus, not required
What we offer
- Ownership of the layer that makes neospaces hard to copy
- Competitive salary plus meaningful equity/upside
- Real proprietary data and real users — your models ship and get used daily
- Modern stack: Claude API, vector search, Inngest, Supabase, Next.js
- Direct line to founders — you sit at the table, not in a ticket queue
- Berlin office, top equipment, flat hierarchies, startup energy
- A clear path to grow as the AI team scales around you If you want your models in production, not in a deck — let's talk.