Position
Aumente sus posibilidades de llegar a la fase de entrevista leyendo la descripción completa del puesto y enviando su solicitud sin demora.
Data Scientist, Retrieval & Agentic Systems (full‑time, Madrid)
Responsibilities
-
- Own and evolve Corestream, Kharon's risk intelligence feed – a dynamic platform that surfaces entity‑level insights and topic‑based risk narratives to customers.
- Redesign and improve Kharon's semantic search capabilities, leveraging modern approaches such as BM25 tuning, semantic distance, Retrieval‑Augmented Generation (RAG), and reranking.
- Integrate Kharon's risk analysis outputs into the search layer, ensuring that graph‑based insights surface meaningfully within search results and sub‑network exploration.
- Drive the evolution of Kharon's narrative generation pipeline – transforming graph and analytical data into compelling, accurate written intelligence for customers.
- Partner with Product and Engineering to take ownership of the Corestream roadmap and push it to its next stage.
- Tweak and optimize metrics that define search quality and narrative relevance.
Qualifications
-
- Proven interest or experience in global security, financial crimes, sanctions, export controls, or related domains.
- 3+ years of experience in a Data Science or AI Engineering role.
- Hands‑on experience with semantic search, including familiarity with BM25, vector search, reranking, and retrieval evaluation.
- Experience with RAG pipelines and working with agentic systems in production.
- Proficiency in Python and SQL.
- Strong product alignment – able to read a roadmap, take ownership of tickets, and drive initiatives forward.
- Independently (bonus): experience with graph‑aware search or integrating knowledge‑graph outputs into retrieval systems.
Company Overview
Kharon operates at the nexus of integral security, helping clients understand and operationalize risk data to make informed decisions. From financial crimes and sanctions to export contro
📌 Data Scientist, Retrieval (Madrid)
🏢 Kharon
📍 Madrid