MLOps engineer at Вериграм — NeverHard

Verigram is an international AI company focused on digital identity and fraud prevention. Our solutions are used by banks, fintechs, and online platforms to onboard users remotely and prevent fraud. And now we are looking for an MLOps Engineer who can take models from research → production → scale. This role is perfect if you love building LLM systems, computer vision pipelines and high-load AI services used by thousands of real users. We build production-ready AI systems that process large-scale data every day: Face biometrics and liveness detection Document verification and KYC/AML Real-time video analytics AI assistants and RAG systems High-load ML services used by thousands of users What you will do: Develop ML services using FastAPI Containerize and deploy models Set up CI/CD Optimize GPU inference and performance Build batch and real-time inference pipelines for LLMs and other models Work with R&D and backend teams Engineering culture: Participate in architecture decisions Code reviews and documentation Contribute to ML engineering best practices Tech stack: Python • PyTorch • FastAPI • PostgreSQL • Redis • Kafka • Docker LangChain • LangGraph • RAG • Vector Search • GPU inference Requirements: Must-have Experience in MLOps / DevOps / Backend 2 - 4 years Experience shipping ML models to production Understanding of the full ML lifecycle Strong level of knowledge of Python, FastAPI, Docker, Linux, SQL Experience in CI / CD Nice to have: Experience with LLMs / RAG / embeddings Inference optimization experience Experience building high-load services Experience in backend development Experience with GPU inference Experience in setting up model monitoring What we offer: Real production ML & LLM projects Access to GPU infrastructure Strong impact on AI product architecture Competitive salary Strong engineering culture, low bureaucracy Join us to build real-world AI products!