Senior AI Engineer — GenAI & Autonomous Agents
Jus Mundi SAS2 months ago
Paris, FranceSenior
Responsibilities
- Own and improve multilingual retrieval, search relevance, ranking, and RAG pipelines over a large legal corpus.
- Design and build autonomous agent systems with orchestration loops, tool and function calling, memory, planning, state management, multi-step reasoning, and error recovery.
- Create rigorous evaluation harnesses, regression suites, benchmarks, and automated tracking for relevance, hallucination, citation accuracy, and model degradation.
- Fine-tune open-source and proprietary models, including with LoRA and QLoRA, and generate synthetic training and evaluation data.
- Deploy AI features to production while managing latency, cost, reliability, graceful degradation, and model failure scenarios.
- Build shared evaluation infrastructure, retrieval components, agent scaffolding, libraries, and engineering patterns for other squads.
- Translate ambiguous business problems into technical roadmaps and ship solutions with measurable impact.
- Mentor engineers through code reviews and RFCs and contribute to interviewing, hiring, onboarding, and technical standards.
Requirements
- Proven experience shipping LLM-powered features to production with APIs or models from OpenAI, Anthropic, Mistral, and Hugging Face.
- Strong expertise in vector databases, retrieval, ranking, RAG architecture, and relevance evaluation.
- Demonstrated hands-on experience designing and shipping agent orchestration systems, using custom architectures or frameworks such as LangGraph, LangChain, LlamaIndex, or AutoGen.
- Deep understanding of evaluation methodology, benchmark design, metric interpretation, automated evaluation pipelines, and the limitations of evaluation methods.
- Demonstrated experience fine-tuning models with LoRA or QLoRA, preparing training data, and generating synthetic data.
- Strong Python skills for AI and backend engineering, plus sufficient TypeScript, Node.js, and React experience for end-to-end product integration.
- Experience with scalable API design, graceful degradation, testability, and reproducibility.
- Ability to collaborate with product, design, and stakeholders, communicate results to non-technical audiences, mentor peers, and take accountability for outcomes.
- Preferred experience includes legal technology or other high-stakes domains, multilingual NLP or retrieval, vLLM, TensorRT-LLM, AWS, GCP, Azure, containerization, and open-source AI contributions or shipped AI applications.
Benefits
- Hybrid work arrangement with remote and on-site work two days per week.
- Five weeks of vacation per year.
- Private health insurance fully covered for the employee and children.
- Company-funded restaurant vouchers covering 60% of weekday meals through Swile.
- Company-wide JM Fest summits.
- Co-financed ClassPass fitness and wellness membership.
- Ability to work from anywhere globally for up to eight weeks per year.
- Relocation assistance to France.
- Support for ongoing training and professional development.
- Inclusive and diverse workplace culture.
Categories
About Jus Mundi SAS
Jus Mundi builds an AI-powered arbitration intelligence platform for law firms, in-house counsel, and researchers, delivered as subscription SaaS. Its products include multilingual legal research (Jus Mundi), a professional selection directory (Jus Connect), and an AI assistant (Jus AI), powered by exclusive data from partnerships with 100+ arbitral institutions. Founded in 2019 and privately held, the company is headquartered in Paris with offices in New York, London, and Singapore.