MWS AI is part of the MTS Web Services ecosystem, where future AI solutions are created. Here we develop modern AI solutions for various fields: from banking to medicine and telecom. We strive for innovation in speech AI, video analytics, and Gen AI: we actively engage in research and publish scientific papers. At MWS AI, you will find a professional community of like-minded people and opportunities for implementing breakthrough projects.
We are looking for an experienced AI Engineer who can transform complex business tasks into working, scalable AI solutions. The focus is on solving real business problems, not on research. Strategic thinking is important to us: the ability to quickly assemble PoCs, validate them on metrics, and iteratively bring MVPs to production.
Main tasks:
- Analysis and decomposition of business tasks
- Design of AI solution architecture together with the team: selection of appropriate patterns
- Implementation of quality metrics (accuracy, relevance, response time), regular analysis of results, hypothesis formulation, iterative improvement of the solution
- Development of AI agents and copilot systems (agent orchestration, memory management, task planning and decomposition, intelligent routing)
- Building RAG systems: selection of indexing strategies, optimal chunking determination, metadata enrichment, implementation of hybrid search and re-ranking
- Versioning of prompts, models, datasets
- Implementation of observability solutions: tracing call chains, monitoring response quality, controlling token costs and response time
- Interaction with a team of diverse specialists (backend/devops/ml/prompt-engineer)
We expect from you:
- 3+ years of applied AI/ML development with a focus on NLP and LLM systems
- Experience with LLM APIs: prompt engineering, context management, streaming
- Practical experience with LLM agents: ReAct pattern, function calling, chain-of-thought reasoning, multi-agent systems
- Mandatory knowledge of LangFlow, n8n or similar
- Deep understanding of RAG: vector and hybrid search, chunking strategies, re-ranking
- Strong Python: clean, typed code, unit tests, asynchronous programming, OOP, Docker
- Knowledge of frameworks: LangChain, LangGraph, CrewAI or similar
- Vector databases: Milvus, Chroma or similar
- LLM observability tools: LangFuse or similar platforms
- Understanding of microservice architecture
- Experience with FastAPI or similar
- Ability to explain complex AI concepts to non-technical stakeholders
- Experience with fine-tuning and model distillation, as well as understanding of modern network architectures will be an advantage
- Practical experience with Qwen, GLM, DeepSeek families and similar
- Understanding of LLM deployment principles: inference servers (vLLM, sglang), scaling, quantization, GPU trade-offs
What we offer:
- 5/2 schedule, partial or full remote work, flexible start of the day.
- Salary + quarterly and annual bonuses.
- Comfortable office a 10-minute walk from Kurskaya metro station.
- Voluntary Medical Insurance (VMI) with dental care, travel insurance, life insurance.
- Participation in conferences and meetups, company-sponsored training.
- Free mobile communication.
- Special offers from MTS partners and friends.
