ВеликобританияMad Devs
Middle Applied AI / LLM Engineer (Python)
1 нед. назад
2300–2800 RUB / мес.ВеликобританияMiddleRemoteB2
pythondjangoceleryredissqlpandaslangchainpostgresql
The role involves owning the AI layer of Mad Devs' products including designing and building LLM pipelines and production-grade Python services, with a focus on real-world GenAI challenges. Mad Devs is a full-stack team developing large-scale IT products across diverse industries, offering remote work outside Russia and Belarus.
About the company
- Mad Devs is a full-stack team specializing in developing and administering large-scale IT products across diverse industries.
- Our projects span Europe, the USA and South East Asia, giving you the opportunity to make a global impact.
- We value flexibility, remote working and professional development, ensuring you thrive in an inspiring environment.
- Join us in revolutionizing the legal industry with AI!
Responsibilities
- Design and build LLM pipelines, AI agents, and RAG-based systems: function/tool calling, structured output (JSON schema).
- Develop production-grade Python services (Django, Celery, Redis).
- Validate product hypotheses using real-world data and analytics.
- Measure and optimize LLM quality, performance, and cost.
- Collaborate closely with product managers and engineers to turn ideas into scalable solutions.
Requirements
- Experience with NLP projects in production, 1+ year.
- Strong experience building production LLM applications.
- Hands-on expertise with RAG, embeddings, vector search, prompt engineering, and LLM evaluation.
- Experience with LangGraph, LangChain, or similar frameworks.
- Solid Python backend background, 2+ years (Django, PostgreSQL, async programming).
- Strong SQL/Pandas skills and a data-driven mindset.
- Language proficiency: English at B2–C1, Russian at C1.
- Location: any, except Russia and Belarus.
Nice to have
- Experience with vector databases (Qdrant, pgvector, etc.).
- Fine-tuning, text classification, or embedding models.
- Kubernetes, gRPC, or Go experience.