A bit about us: Tabnine is redefining how software gets built. Trusted by many enterprise-grade companies, we build AI-first developer experiences powered by state-of-the-art coding agents and code reasoning models. With support for over 30 programming languages and 15+ IDEs, our platform is pushing the limits of LLM-based software engineering, enabling teams to design, write, review, and ship code faster than ever. We're committed to advancing code-native AI models, multi-agent systems, agent orchestration frameworks, and autonomous dev tooling to empower developers at every step of the software lifecycle.

We're growing rapidly, and our team is passionate about pushing AI engineering to new heights, solving complex problems in agent orchestration, model integration, and building production AI systems at scale.

About the Role: As an AI Engineer, you'll build and ship intelligent coding systems that combine the best foundation models with custom agents, tool integration, and sophisticated orchestration. You'll be hands-on with LLMs daily, selecting the right model for each task, building agentic workflows, designing sophisticated context tools, and crafting systems that deeply understand code. Your work will directly impact how millions of developers create software.

What You'll Do:

• Build AI Agent Systems: Design and implement multi-agent architectures, handling complex coding workflows from planning through execution.
• Master Model Selection & Integration: Work with multiple LLM providers (OpenAI, Anthropic, open-source models) - understanding their strengths, context windows, reasoning capabilities, and cost trade-offs to choose the right model for each use case.
• Work with Open-Source Models: Deploy and optimize open-source models on GPUs for specific use cases - balancing performance, cost, and privacy requirements.
• Create Tool Ecosystems: Develop MCP (Model Context Protocol) servers and custom tools that extend LLM capabilities, integrating with IDEs, version control systems, testing frameworks, and development environments.
• Design Agentic Workflows: Implement RAG pipelines, function calling, tool use, memory systems, and multi-turn reasoning flows that make coding agents truly useful and context-aware.
• Build Evaluation Systems: Create both offline evaluation frameworks (benchmark datasets, automated testing, and regression detection) and online evaluation systems (A/B testing, user feedback loops, and production metrics) to measure and improve agent quality continuously.
• Optimize for Production: Take AI systems from prototype to production - handling prompt engineering, caching strategies, error recovery, latency optimization, and cost management at scale.
• Experiment & Iterate: Rapidly prototype new agent capabilities, test different model approaches, and use evaluation frameworks to measure and improve agent performance.
• Ship AI Features: Collaborate with product and engineering teams to deliver AI-powered developer tools that users love and rely on daily.

What We're Looking For:

• 2+ years building with LLMs and AI systems - you've shipped products using LLMs to production
• 4+ years of professional experience in software engineering, data science, or machine learning, with the last 2+ years focused on building and deploying AI/LLM systems to production
• Hands-on agent development experience
• Deep model knowledge - you understand the differences between LLMs and can articulate when to use each
• Evaluation mindset - experience building offline benchmark suites and online evaluation systems to measure AI system performance in production
• Production mindset - you think about latency, cost, reliability, and user experience, not just what's technically possible
• Strong coding skills - proficient in Python; comfortable with async programming, API design, and working with complex codebases
• Passion for developer tools - you care deeply about making developers more productive and love experimenting with AI coding assistants

Nice to have:

• Experience running open-source models on GPUs
• Knowledge of GPU optimization, batching strategies, and inference acceleration techniques
• Built MCP servers or contributed to the MCP ecosystem
• Experience with prompt engineering frameworks and evaluation systems
• Familiarity with model quantization for efficient inference
• Experience with vector databases and semantic search
• Built AI coding assistants, code analysis tools, or developer-focused AI products
• Understanding of software engineering workflows (CI/CD, code review, testing) and how to augment them with AI
• Experience designing A/B tests and analyzing user behavior metrics for AI features
• Contributions to open-source AI frameworks or developer tools
• Experience in fast-paced startup environments shipping AI products