Full time · Middle · Remote
Salary: from 1 800 USD
Job description
We are looking for an AI/LLM Engineer who will design and develop production systems based on LLM.
❕We need not a classic Data Scientist or a specialist who only deals with prompts. We are looking for an engineer who knows how to assemble an LLM system entirely: from prompt/context logic and agent scenarios to backend services, integrations, testing, and deployment to production.
📍 MAIN STACK — Go + Python.
⏺ Go — we use for production backend, services, and high-load business logic.
⏺ Python — for LLM infrastructure, AI components, prototyping, and experiments.
💎 Responsibilities:
- Design and develop LLM/AI systems and AI agents;
- Build multi-step agentic workflows: routing, tool calling, state management;
- Work with Prompt Engineering and Context Engineering;
- Manage context, memory, and dialogue state;
- Implement structured output and function/tool calling;
- Build RAG pipelines: chunking, embeddings, retrieval, reranking;
- Integrate LLM with APIs, databases, and internal services;
- Develop backend services in Go and AI components in Python;
- Research new models and approaches, quickly test hypotheses;
- Select models considering quality, latency, and cost;
- Build a quality assessment system: evals, regression tests, golden datasets, LLM-as-a-judge;
- Analyze logs and traces, improve agent quality and stability;
- Implement retries, fallbacks, timeouts, and error handling;
- Take solutions from prototype to stable production.
💯What we expect:
Go:
- production services;
- REST/gRPC API;
- concurrency;
- database interaction;
- microservice architecture.
Python:
- AI services and backend;
- async;
- API;
- data processing;
- rapid prototyping.
🔥 Also important:
- Practical experience in developing production applications with LLM;
- Understanding of modern AI system architecture:
prompt → context → tools → agent → state → evaluation;
- Experience in Prompt Engineering: system prompts, few-shot, prompt chaining, structured output;
- Experience in Context Engineering: memory, summarization, dynamic context assembly;
- Experience in creating AI agents and agentic workflows;
- Working with LLM APIs and understanding the specifics of different models;
- Understanding of RAG: embeddings, vector search, retrieval, reranking, hybrid search;
- Understanding of evals, regression testing, and production log analysis;
- Experience with PostgreSQL, MongoDB, Redis, Docker, Git, CI/CD;
- Understanding of production approaches: observability, scalability, latency, retries, fallbacks, and cost control.
🌟 Will be a plus:
- Multi-agent systems;
- LangGraph;
- MCP;
- Langfuse / LangSmith;
- Elasticsearch / OpenSearch;
- Open-source LLM: Qwen, Llama, DeepSeek, etc.;
- vLLM / Ollama;
- Kafka / RabbitMQ;
- Kubernetes;
- Optimization of token usage, latency, and LLM cost;
- Guardrails and protection against prompt injection;
- High-load systems.
‼️ IMPORTANT
We are looking for an engineer who perceives LLM not as a simple API, but as a component of a large production system.
It is important to be able to design all the logic around the model: what data to pass into the context, which agent or tool to call, how to check the result, handle errors, and determine the next step.
We are looking for a software engineer with strong AI/LLM expertise, who can not only write a good prompt but also bring the entire system to a stable production solution.
