Senior AI Engineer

The OpportunityIf you've worked alongside hardware teams, you know the damage that results from a missed change request or critical context that was never relayed to the right person. Reflow exists to close that gap. We're building the first AI-powered platform built for hardware product development, one that listens across the tools teams already use, maintains a structured picture of every program, and proactively coordinates across disciplines when things inevitably change.This is an early role for a hands-on AI engineer to design and build the agents at the core of our platform, backed by a parent company with deep roots in engineering and manufacturing. You'll own the AI systems that set our product apart: agents that understand hardware workflows, anticipate problems, and take action on behalf of engineering teams.Who We're Looking ForWe're looking for an AI engineer who works equally well in applied research and production software. You've shipped LLM-powered agent systems to real users, you have strong intuitions about prompt engineering, tool use, and orchestration patterns, and you keep up with a field that changes fast. You're comfortable evaluating new frameworks, models, and techniques on short cycles and making pragmatic build-vs.-adopt decisions as the landscape shifts.We currently use LangChain's DeepAgent framework for our agent orchestration, but every layer of our AI stack is open to change. The right person for this role understands how agent systems work under the hood, not just one framework's abstractions, and can adapt as better tools and approaches emerge.This senior engineer will work alongside our engineering team and head of product to build, optimize, and operate the AI agents that give users proactive coordination, risk surfacing, status summaries, and AI-generated deliverables. The role is hands-on. You'll write code daily while contributing to AI architecture decisions and helping define how we evaluate and evolve our agent capabilities over time.What You'll DoYour day-to-day responsibilities will include: • Designing, building, and iterating on LLM-powered agents that coordinate across engineering disciplines, surface project risks, and generate structured deliverables (proposals, SOWs, status reports) • Owning the agent orchestration layer (currently LangChain DeepAgent) and continuously evaluating whether to extend, replace, or supplement it as new frameworks and patterns emerge • Implementing robust tool-use patterns that connect agents to external systems (project management tools, CAD/PLM platforms, communication channels) via APIs and integrations • Designing and tuning prompts, chains, and retrieval strategies to maximize agent reliability, accuracy, and usefulness across diverse hardware project contexts • Building evaluation and observability infrastructure for agent performance, including tracing, cost tracking, latency monitoring, and automated quality benchmarks • Developing streaming agent interfaces that surface real-time progress, reasoning transparency, and proactive alerts to end users • Staying current with rapid advances in LLMs, agent frameworks, and related tooling, and translating that awareness into actionable recommendations for the team • Collaborating with frontend engineers on the UX of AI-powered features and with backend engineers on data pipelines and API design • Contributing to AI architecture decisions, code reviews, and engineering best practicesTechnical RequirementsMust Have: • 5+ years of production software engineering experience, with 2+ years focused on bringing LLM-based applications or agent systems to market • Demonstrated proficiency using AI coding tools (Cursor, Copilot, Claude, etc.) to accelerate development • Hands-on experience building and deploying agentic systems using frameworks such as LangChain/LangGraph, CrewAI, AutoGen, or custom orchestration • Strong understanding of LLM fundamentals: prompt engineering, function/tool calling, retrieval-augmented generation (RAG), context window management, and token economics • Proficiency with Python in production environments • Experience integrating LLM-powered features with external APIs, databases, and third-party tools • Experience designing and operating background job / async task pipelines (Celery, RQ, Temporal, or similar) for long-running agent runs and reliable retries • Experience building multi-agent systems with planning, delegation, and inter-agent communication patterns • Demonstrated ability to evaluate and adopt new AI tools and frameworks quickly, with a track record of staying ahead of a fast-moving field • Strong software engineering fundamentals: clean architecture, testing, version control, and code review practices • Ability to balance rapid experimentation with production-grade reliabilityHighly Valuable: • Direct experience with LangChain's DeepAgent or LangGraph for multi-step agent orchestration • Background in evaluation frameworks for LLM outputs (aut