Company: Agentic Lab
Employment Type: full-time
Salary: 200000–500000 RUB per month
Format: hybrid / remote with Russian Federation employment registration
Development and implementation of industry-grade AI solutions on top of our proprietary platform — agentic workflows, multi-agent systems, RAG, and LLM pipelines, from prototype to production for enterprise clients.
Responsibilities
- Develop industry-grade AI solutions on top of the platform: agentic workflows, tool calling, RAG, memory, and human-in-the-loop loops
- Design multi-agent systems without strict dependency on a specific framework or SDK — agent interaction logic, state, memory, routing, and error handling
- Evaluate and optimize LLM pipeline quality — Langfuse, custom evals and quality metrics, latency, cost, and reliability
- Integrate AI systems into clients' existing IT landscape and deliver solutions from prototype to production
Requirements
- Relevant experience in AI/ML, backend, or software engineering
- For Mid/Senior level: proven experience in designing and launching LLM-based systems in production, as well as deploying and optimizing open-weight LLMs
- Proficient Python and experience in developing production systems
- Proficient work with PostgreSQL and vector databases; ClickHouse/Greenplum and Redis are a plus
- Business acumen — understanding of the business problem and responsibility for the solution's value, not just technical implementation
- Ownership and proactivity — willingness to independently lead a task from its inception to a working result
Conditions
- Competitive market compensation
- Real enterprise tasks: ontological platform, multi-agent systems, corporate integrations, and AI implementation in real business processes
- Opportunity to influence not only client solutions but also the platform's architecture and development
- Hybrid / remote with Russian Federation employment registration
Skills
- Python
- PostgreSQL
- vector databases
- ClickHouse
- Greenplum
- Redis
- LLM
- RAG
- agentic workflows
- tool calling
- multi-agent systems
- Langfuse
- evals
- open-weight LLM
