Informacje o stanowisku

DealSumm is a California-based startup operating in the American and European legal market. Our product is a SaaS platform that analyzes and manages legal contracts and already serves the largest players in the market. Our core technology is Natural Language Processing and Large Language Models (LLMs).

We are looking for an experienced AI Agents & Databricks Specialist to join our R&D team and own the design, development, and ongoing operations of a production-grade prediction and retrieval pipeline. In this role, you will build and maintain systems that combine agents, RAG, vector search, hybrid search, re-ranking, evaluation/judging, and fine-tuning, running at scale on Databricks.

Requirements

• BSc in Computer Science / Computer Engineering (or equivalent experience)
• 4+ years professional software engineering experience, with 2+ years focused on ML/LLM systems or search/retrieval pipelines
• Strong Python engineering skills (production quality code, testing, performance)
• Hands-on experience with Databricks (jobs/workflows, Delta Lake, Spark, MLflow, model/pipeline operations) — must
• Proven experience with RAG and retrieval systems: embeddings, vector databases/vector search, chunking strategies, evaluation
• Experience with hybrid search and relevance tuning (BM25 + semantic, query rewriting, re-ranking)
• Experience with re-rankers and ranking approaches (cross-encoders, LLM-based reranking, learned ranking strategies)
• Experience building evaluation/judge frameworks (human + automated eval, LLM-as-judge, quality metrics, A/B testing)
• Familiarity with LangChain (or equivalent orchestration frameworks) and agent/tooling patterns
• Strong understanding of LLM fine-tuning concepts and practical decision-making (when to fine-tune vs. RAG vs. prompting)
• Experience with containers (Docker) and Linux — advantage
• AWS Cloud experience — advantage
• Excellent problem-solving skills, ownership mindset, and strong communication
• Full proficiency in English
• This position is open to both men and women.

DealSumm is a California-based startup operating in the American and European legal market. Our product is a SaaS platform that analyzes and manages legal contracts and already serves the largest players in the market. Our core technology is Natural Language Processing and Large Language Models (LLMs).

We are looking for an experienced AI Agents & Databricks Specialist to join our R&D team and own the design, development, and ongoing operations of a production-grade prediction and retrieval pipeline. In this role, you will build and maintain systems that combine agents, RAG, vector search, hybrid search, re-ranking, evaluation/judging, and fine-tuning, running at scale on Databricks.

,[Design, build, and maintain end-to-end LLM/agent systems in production (agents + tools + workflows), Own and evolve our Databricks-based infrastructure for data processing, feature engineering, and model/pipeline execution, Implement robust RAG pipelines, including ingestion, chunking, embeddings, vector indexing, hybrid retrieval, and re-ranking, Develop and optimize vector search and hybrid search (semantic + keyword), including relevance tuning and latency optimization, Build evaluation frameworks (offline + online), including LLM-as-a-judge, golden datasets, regression tests, and quality dashboards, Lead experimentation and rollout of fine-tuning (when needed), prompt strategies, and model selection, Ensure reliability: monitoring, alerting, cost control, performance, and safe deployment practices, Work closely with the VP R&D and cross-functional stakeholders to translate product needs into scalable AI systems, Research and adopt new tools, methods, and best practices in the fast-evolving GenAI ecosystem Requirements: Spark, Python, Machine learning, Databricks, Docker, Azure

Praca Wrocław

Wrocław - Oferty pracy w okolicznych lokalizacjach