We are seeking a Senior AI Engineer to architect and scale the intelligent brain powering our virtual personal assistant. In this role, you will own the technical roadmap for our LLM pipelines, driving the transition from basic prompting to complex, multi-agent workflows.
You will design production-grade systems that accurately parse, reason over, and execute tasks derived from unstructured household data - like photos, school PDFs, and fragmented calendar updates - while maintaining ultra-high accuracy and data privacy for busy families.
Key Responsibilities
- Architect Multi-Agent Systems: Design, build, and deploy reliable agentic workflows and function-calling infrastructures that execute multi-step household tasks (e.g., cross-referencing school PDFs to populate multiple family calendars and trigger grocery orders).
- Advanced Multimodal Extraction: Develop high-accuracy document parsing and vision-processing pipelines to extract structured timelines, meal requests, and tasks from messy physical or digital documents. Converse with the user through multiple channels (voice, chat).
- Contextual Memory & RAG Systems: Architect scalable, low-latency long-term memory systems (using advanced RAG and graph-based vector structures) to maintain strict data separation and hyper-personalized context across multi-user household profiles.
- Infrastructure & Optimization: Optimize inference costs, reduce system latency, and implement robust evaluation frameworks to continuously test prompt modifications, model upgrades, and routing decisions.
- Technical Leadership: Define the engineering standards for AI application development, mentor mid-level engineers, and establish strict data security guardrails to safeguard sensitive family data.
Qualifications & Technical Skills
- Core Experience: 5+ years of production software engineering experience, with at least 2–3 years spent building and scaling production-grade LLM applications.
- Production AI Tooling: Advanced mastery of LangGraph, LangChain, or custom-built orchestration frameworks, alongside direct experience managing structured outputs (e.g., Pydantic parsing) from commercial and open-source foundation models.
- Data Architecture: Extensive expertise using vector databases (e.g., Pinecone, Milvus, pgvector), traditional relational databases, and designing real-time, asynchronous data pipelines.
- Production Operations: Heavy experience deploying containerized services on cloud infrastructure (AWS or GCP), managing CI/CD pipelines for AI applications, and setting up logging/observability platforms (e.g., LangSmith, Arize).
- Location & Hours: Fully remote within the United States. Must be able to maintain core working hours aligned with Eastern Standard Time (EST).