• Build foundation models and generative AI tools alongside a team of technologists. • Design and build agentic workflows — multi-agent orchestration (e.g., CrewAI, LangGraph, AutoGen), tool use, multi-step planning, and human-in-the-loop checkpoints — to automate complex engineering tasks. • Establish evaluation, guardrails, and failure-mode analysis for agent systems to ensure they are safe, reliable, and production-ready. • Develop scalable data pipelines for diverse data sources used in production ML systems, including BIM, CAD, and infrastructure design data. • Work with large-scale, multi-modal datasets — including text and geometric data — to design novel preprocessing, augmentation, analysis, and content understanding. • Transform unstructured infrastructure and design data into representations suitable for machine learning. • Lead cross-functional collaboration with ML Research Scientists and Engineers to align data formats with downstream training and fine-tuning of LLMs. • Apply deduplication, normalization, and validation techniques to ensure high-quality data in production environments. • Architect and optimize pipelines for scalability, reproducibility, and cloud deployment. • Mentor junior engineers and provide technical guidance on complex data challenges. • Drive technical decision-making and influence best practices across the team. • Perform requirements analysis with senior stakeholders, ensuring technical solutions meet both immediate project goals and long-term research objectives. • Communicate findings and technical insights through quantitative analysis, visualizations, and clear documentation. • Contribute to agile workflows, ensuring flexibility and responsiveness to evolving project needs. • Participate in technical planning and roadmap development.

• Lead agentic AI development, including multi-agent orchestration patterns, agent-to-agent protocols, and reliable tool use at production scale • Own prompt engineering and evaluation workflows including structured outputs, hallucination reduction, and behavioral consistency • Build and own MCP server infrastructure that exposes backup data to AI agents via the Model Context Protocol, enforcing tenant-aware RBAC, query constraints, and safe tool boundaries • Define AI quality benchmarks for retrieval relevance, summarization accuracy, and agent reliability, and drive systematic improvements through eval-driven iteration • Champion security and safety in AI systems, including adversarial prompt hardening, jailbreak resistance, data boundary enforcement, and OWASP LLM Top 10 awareness • Tune AI workflows for performance, cost, latency, and observability across billions of documents in global regions • Mentor engineers on the team, raise the technical bar, and contribute to architecture reviews and design decisions

• Own the AI roadmap. • Set direction across our AI surfaces: voice and computer-use agents that work payers and provider portals, email triage, and the models that extract and structure data from clinical and legal documents. • Push AI deeper. • Take our systems from assisting people, to making decisions, to running whole service requests, and into licensing, enrollment, monitoring, and the customer-facing product. • Make it production-grade. • Build the evals, observability, and human-in-the-loop systems that let us ship models and agents we'll stand behind and back with guarantees. • Lead the data platform. • Own the pipelines, datasets, and tooling behind the models. Our real-world data is our biggest advantage, so make it fast to use and trustworthy to build on. • Build the org. • Hire and grow managers and senior engineers, set the structure, and hold the bar as we scale. Stay close enough to the work to earn their trust in a design review.

• Design & Develop Agentic Systems: Build intelligent agents capable of autonomous planning, reasoning, and task execution, often using LLMs (e.g., GPT-class, LLaMA), multi-modal models, and autonomous workflows • Orchestration & Frameworks: Implement agent orchestration using frameworks like LangChain, AutoGen, CrewAI, Semantic Kernel, or custom solutions • Retrieval-Augmented Generation (RAG): Design and optimize RAG pipelines for enhanced reasoning with external knowledge, including document ingestion, chunking, embeddings, vector stores, and retrieval ranking • Tool & Memory Integration: Develop agents that call APIs, databases, and other tools, maintain memory, and adapt based on outcomes • Evaluation & Monitoring: Create evaluation frameworks for accuracy, grounding, latency, and cost; build observability for agent behavior and failure modes • Model Adaptation: Fine-tune or adapt foundation models (e.g., via LoRA, adapters) for domain-specific use cases • Production Deployment: Deploy GenAI/agentic systems in cloud-native environments with CI/CD, versioning, and runtime safeguards • Cross-Functional Collaboration: Work with data scientists, ML engineers, product teams, and governance/compliance stakeholders