SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Built In is a venture-backed platform for tech professionals globally, with millions of monthly visitors and 1,800 customers ranging from startups to Fortune 100 companies. The company has raised $30 million in funding and is expanding its geographic footprint, product suite, and technology capabilities.
You will join as an Applied AI Engineer to design and build autonomous systems powering Built In's product suite. Your core mission is to architect the Agentic Action Layer that transforms analytical insights into intelligent outcomes. You will create multi-agent systems that take supervised action across frontier AI models (ChatGPT, Google Gemini, Perplexity), operating with a human-in-the-loop mindset to compress traditionally team-sized efforts into individual throughput without sacrificing reliability.
Key responsibilities include:
• Architect production-grade agentic workflows using Reflection, Self-Correction, Planning, and Multi-Agent Orchestration patterns to power AI-driven product capabilities.
• Own the full lifecycle from rapid prototype to production deployment, building agent systems with real-world constraints: token budgets, latency targets, graceful degradation, and cost observability.
• Develop systems that synthesize outputs from multiple frontier models to identify sentiment discrepancies, detect brand hallucinations, and surface actionable intelligence for customers.
• Design and maintain serverless backend services using Python and AWS Lambda, ensuring agents perform efficiently at scale with proper observability.
• Build LLM-as-a-Judge evaluation frameworks, trace-based testing pipelines, and quality feedback loops to keep agent output reliable as models and prompts evolve.
• Collaborate with Software Engineers and Product Managers to integrate agentic behaviors into production frameworks, ensuring AI systems are stateful, observable, and resilient.
Requirements:
• Deep experience designing and shipping autonomous systems using patterns like ReAct, Planning, Reflection, and Multi-Agent Orchestration. You must have moved agents from demo to production and dealt with failure handling, state management, and non-deterministic outputs.
• Expert-level implementation experience with frontier model APIs from OpenAI, Google, and Perplexity, including function calling, structured outputs, and streaming.
• Demonstrated power-user proficiency with agentic developer tools (Claude Code, Cursor, or custom agentic CLI workflows) for architecture, refactoring, testing, and shipping—not just autocomplete.
• Demonstrated ability to manage multiple autonomous agents across a codebase: writing, testing, and auditing code with minimal manual intervention.
• Strong proficiency in Python for AI orchestration, plus working knowledge of C# / .NET. The core platform runs on .NET, and you must integrate agent systems with it.
• Practical experience deploying and operating serverless architectures on AWS (Lambda, Step Functions, EventBridge) for high-frequency agentic workloads.
• Experience building custom evaluation frameworks (LLM-as-a-Judge, trace-based testing, regression suites) to maintain and improve agent output quality over time. You must understand that the hardest part of production agents is knowing when they're wrong.
