Просмотр вакансии

AdQuantum — a strategic marketing partner for successful businesses in the Web & App industry. We offer a full-scale creative production studio and extensive resources that help mobile apps and web projects discover growth opportunities and achieve high profitability at every stage. Since 2015, we’ve delivered positive ROI for over 200 brands, including Playrix, Madbox, Verv, Flo, Playgendary, Octa FX, and many others.

Responsibilities

• Design and deploy LLM-powered agents to automate internal business processes

Including data analysis, report generation, inbound request processing, and task routing across teams.

• Build integrations between AI models and the company's internal systems

Connect CRM platforms, task trackers, messaging tools, and advertising platforms through APIs and MCP servers.

• Develop RAG (Retrieval-Augmented Generation) systems and internal knowledge search tools

Including document vectorization, retriever configuration, and response quality evaluation.

• Create AI-powered tools for agency teams, including media buying, analytics, and creative departments

Such as automated campaign analysis, creative brief generation, and performance reporting.

• Design prompt strategies, evaluate output quality, and optimize cost and latency

Including fine-tuning approaches, context caching, and selecting the most appropriate model for each use case.

• Participate in the technical design of new LLM-based products

From concept to production: architecture, technology stack selection, deployment, and monitoring.

• Proven experience delivering production-grade LLM solutions

You should be able to provide specific examples, including the business problem solved, system scale, traffic/load characteristics, and measurable outcomes. Academic projects, demos, and prototypes will not be considered.

• Strong hands-on experience with Anthropic Claude API and/or OpenAI API

Including function calling, tool usage, streaming, context window management, and error handling.

• Experience building AI agent systems

Using frameworks such as LangChain, LlamaIndex, AutoGen, or custom orchestration solutions. Practical experience is more important than knowledge of any specific framework.

• Production-level Python skills

Strong understanding of async/await, API integrations, data structures, and a solid grasp of deployment processes and CI/CD practices.

• Deep understanding of LLM limitations and trade-offs, including hallucinations, performance degradation with long context windows, token cost optimization, latency challenges, and practical mitigation strategies.