ML Architect: AI Context & Retrieval, Remote + ESOPs

Unabyss; Województwo mazowieckie; Remote; PLN 244,000 - 390,000; Full time

Benefits offered by this job

Fully remote

ESOPs planned

Sports benefits

Direct collaboration with CTO

Job summary

Unabyss is building a universal context layer for AI. We seek an Applied Researcher to design and implement a memory system, data pipelines, and retrieval mechanisms with privacy controls.

You will prototype architectures, compare approaches, and collaborate directly with the CTO to shape the product's core technical direction. You should have 2+ years in ML/NLP/LLM production, strong Python, and deep retrieval knowledge.

Qualifications

• 2+ years of professional experience building ML/NLP/LLM systems in production.
• Strong Python proficiency for research and production tasks.
• Deep understanding of retrieval architectures beyond basic RAG - hybrid search, structured extraction, re-ranking, query decomposition.
• Comfort with rapid prototyping - building POCs, running experiments, measuring results, iterating fast.
• Experience using AI coding tools as part of the workflow.
• Strong algorithmic thinking and ability to reason about trade-offs (precision vs recall, cost vs accuracy).
• Experience designing and evaluating ML pipelines end-to-end.
• Ability to work independently and drive research-to-production cycles.
• Fluent English (written and spoken).

Responsibilities

• Design and build a knowledge/memory system that structures user context for precise retrieval.
• Build data processing pipelines to enrich, classify, and structure information as it flows in.
• Design retrieval mechanisms that are cheap, accurate, and traceable to root sources.
• Implement permission-aware access control at the retrieval layer with per-app consent boundaries.
• Evaluate and compare architectural approaches through rapid prototyping and benchmarking.
• Design prompt engineering strategies and LLM pipeline architectures using commercially available models.

Skills

ML/NLP/LLM production

Python

Retrieval architectures

Prototyping & experiments

AI coding tools

Algorithmic thinking

ML pipelines

Independence & decision making

English proficiency

Job description

Unabyss is building a universal context layer for AI. We seek an Applied Researcher to design and implement a memory system, data pipelines, and retrieval mechanisms with privacy controls.

You will prototype architectures, compare approaches, and collaborate directly with the CTO to shape the product's core technical direction. You should have 2+ years in ML/NLP/LLM production, strong Python, and deep retrieval knowledge.