AI engineer

Описание

Lexroom is building the AI platform where legal work gets done. In just over two years, we’ve grown from zero to €25M ARR, and today more than 20,000 legal professionals across Italy, Spain, Germany, France, and Bulgaria rely on Lexroom. We’re backed by $73M from Left Lane Capital, Base10 Partners, and Eurazeo, and we’re a team of around 300 builders who move fast, say things straight, and own each outcome.

Задачи

• Design and build agent-driven and RAG-based AI systems to solve complex legal tasks with precision and efficiency.
• Fine-tune and post-train LLMs using cutting-edge supervised and reinforcement learning techniques to raise response quality.
• Develop and optimize large-scale search pipelines for LLM-powered agents.
• Implement and refine NLP algorithms to extract, structure, and validate information from legal documents.
• Run end-to-end evaluations to pinpoint bottlenecks, measure performance, and guide continuous improvements.
• Work closely with legal experts to design evaluation protocols and build reliable, domain-specific datasets that strengthen every layer of the stack.
• Maintain technical excellence by following established best practices, coding standards, and rigorous code-review processes.

Требования

• Proven expertise in building and managing robust ML pipelines.
• Strong Python skills with hands-on experience in LLM research and deep learning systems.
• Knowledge of search technologies such as full-text search, vector databases, or other large-scale retrieval systems.
• Collaborative mindset and readiness to work with legal experts, researchers, and engineers to keep improving the product.
• Up-to-date with LLM innovation, ideally with experience in pre-training or deploying large-scale models in real-world scenarios.
• Fluency in English, both written and spoken Будет плюсом: Familiarity with MLOps and integrating ML models into production backends (a strong plus); knowledge of Italian is a plus

Условия

• Annual Salary Range: €70,000 - €100,000
• Equity Package: Align your growth with the company’s long-term success.
• Hybrid Work: Flexibility to balance remote work with in-person collaboration at our Barcelona office.
• Health Coverage: Comprehensive insurance to keep you secure and focused.
• Meal Vouchers: Daily support for your lunch breaks.
• Performance Bonuses: Rewards directly tied to measurable results.
• Team Retreats & Off-sites: Dedicated time to connect, collaborate, and spark bold ideas.

• Proven expertise in building and managing robust ML pipelines.
• Strong Python skills with hands-on experience in LLM research and deep learning systems.
• Knowledge of search technologies such as full-text search, vector databases, or other large-scale retrieval systems.
• Collaborative mindset and readiness to work with legal experts, researchers, and engineers to keep improving the product.
• Up-to-date with LLM innovation, ideally with experience in pre-training or deploying large-scale models in real-world scenarios.
• Fluency in English, both written and spoken Будет плюсом: Familiarity with MLOps and integrating ML models into production backends (a strong plus); knowledge of Italian is a plus

• Annual Salary Range: €70,000 - €100,000
• Equity Package: Align your growth with the company’s long-term success.
• Hybrid Work: Flexibility to balance remote work with in-person collaboration at our Barcelona office.
• Health Coverage: Comprehensive insurance to keep you secure and focused.
• Meal Vouchers: Daily support for your lunch breaks.
• Performance Bonuses: Rewards directly tied to measurable results.
• Team Retreats & Off-sites: Dedicated time to connect, collaborate, and spark bold ideas.

Lexroom is building the AI platform where legal work gets done. In just over two years, we’ve grown from zero to €25M ARR, and today more than 20,000 legal professionals across Italy, Spain, Germany, France, and Bulgaria rely on Lexroom. We’re backed by $73M from Left Lane Capital, Base10 Partners, and Eurazeo, and we’re a team of around 300 builders who move fast, say things straight, and own each outcome.