Staff Applied AI Engineer

🚀 About Us We’re Bolt.new by StackBlitz! We’re the team that brought you WebContainers, the first-of-its-kind technology that made it possible to run Node.js right inside your browser. That breakthrough kicked off our journey in 2019, and it’s what powers the blazing-fast online IDE used by over 1 million developers every month. But we didn’t stop there. We doubled down on everything we learned and built Bolt.new — the fastest way to go from idea to production without writing traditional code. It’s a next-gen, AI-powered app builder that helps you create, edit, and deploy full-stack web and mobile apps instantly, right in your browser. No installs. No setup. Just smart automation and instant dev environments that let you move at the speed of thought. We’re a fully remote team, globally distributed, deeply collaborative, and seriously passionate about building the future of software development. This is your chance to join a small team with a big vision. If you love shipping fast, solving real problems, and pushing the boundaries of what’s possible, we’d love to meet you. ✨ About This Opportunity As a Staff Engineer on the AI team, you'll lead the technical direction of the AI agents that turn natural language into production-ready applications. This means shaping how we work with LLMs to solve our hardest problems: maintaining context across large codebases, orchestrating multi-step workflows that feel intuitive, and handling everything from simple UI tweaks to complex architectural decisions. This isn't API integration work. You'll define the patterns and systems that govern how AI reasons about and generates full-stack applications, driving initiatives across multiple teams and influencing our broader AI strategy. Your work shapes the experience of millions of users building real products with Bolt every day. 🛠️ How You'll Contribute Define AI Agent Architecture: Set technical direction for how agents manage context, orchestrate workflows, and scale. Lead multi-model strategy: Build evals and selection criteria across providers (OpenAI, Anthropic, Google); partner with provider teams to test new capabilities. Build tool-use & workflow foundations: Design safe, reliable interfaces for tool calling (search, queries, domain actions); evaluate frameworks (e.g., Vercel AI SDK, LangGraph) and set org best practices. Drive cross-team execution: Align product/design/engineering, resolve tradeoffs, and mentor engineers to raise AI engineering standards. Establish data & evaluation standards: Own dataset methodology and the eval harness; turn failure modes and conversation insights into measurable improvements. Drive Research and Innovation: Experiment with prompting, context handling, and post-training; share learnings externally when appropriate. 💡 Qualifications Deep LLM experience: Built and scaled production LLM systems; strong grasp of capabilities, limits, and emergent behavior. Prompt engineering: Sets best practices and mentors across models and use cases. Software engineering: Strong fundamentals; designs scalable systems and makes pragmatic architectural calls. Strategic execution: Drives ambiguous, high-scope work end to end; influences across teams. Systems thinking: Spots process/communication/technical debt and improves team velocity. Model & agent literacy: Tracks coding-agent/LLM advances; understands model tradeoffs and the agent lifecycle. Data-driven leadership: Builds data collection + eval harnesses; turns insights into measurable improvements. Strong verbal and written English communication skills are required, as this role involves frequent collaboration with team members, stakeholders, and customers where English is the primary working language. 🎯 Bonus Points Fine-tuning and alignment o f: LLMs (SFT, RLHF/RLAIF, DPO/ORPO) Machine Learning Background: Understanding of ML fundamentals and experience with model evaluation metrics. Open Source Contributions: Experience contributing to or maintaining open-sou…