Guardian is undergoing a transformation to become a modern insurance company, and the Data & AI team, specifically the Data Science Lab (DSL), is tasked with creating business value from data and analytic products to drive revenue growth, risk management, and customer experience. This role will focus on developing AI/ML solutions for Document Intelligence, Intelligent Knowledge Retrieval, and AI Agents to enhance operational efficiencies and financial performance.
Requirements
- 7+ years building and shipping ML/AI solutions, including 3+ years leading data science teams; experience managing a team of practitioners.
- Document Intelligence (IDP) expertise: OCR/layout understanding, document classification, entity & table extraction, summarization; comfortable with modern document models and OCR services.
- LLMs & GenAI: prompt design and tuning (incl. PEFT/fine-tuning), tool/function calling, safety guardrails; hands-on with retrieval-augmented generation (RAG), vector search, chunking, and grounding over enterprise content.
- AI Bots & Agentic Automation: experience building task-oriented agents and bots for intake, triage, and workflow orchestration with human-in-the-loop controls.
- Strong Python plus PyTorch and/or TensorFlow; solid algorithms background and proficiency in data/ML engineering patterns (e.g., Spark or Ray) and GPU/distributed computing.
- MLOps proficiency: CI/CD for models and prompts, experiment tracking (e.g., MLflow/W&B), model/feature/embedding stores, observability/monitoring, A/B and champion-challenger testing.
- Evaluation mindset end-to-end: document accuracy, retrieval quality, grounding/hallucination, safety, latency/cost, tied to business KPIs.
Responsibilities
- Lead the end-to-end lifecycle—from discovery and pilot to enterprise rollout—of IDP use cases (classification, OCR/layout understanding, entity/table extraction, summarization) across underwriting, claims, and customer service.
- Design and ship retrieval-augmented solutions (RAG/grounded generation, vector search, conditional routing) for policy, clinical, and knowledge assets.
- Build agentic automations and AI bots for intake, triage, and workflow orchestration; integrate with core systems and human-in-the-loop reviews.
- Evaluate and adapt advances in deep learning, LLMs, multimodal document models, and tool-use/agents to high-impact insurance scenarios.
- Implement multi-layer evaluation (document accuracy, retrieval quality, grounding/hallucination, safety, latency/cost) with automated regression tests.
- Optimize compute and inference (GPU/distributed, caching/batching, model routing) for cost and performance.
- Embed Responsible AI by design: privacy/PHI/PII safeguards, explainability, bias monitoring, content safety, and model risk documentation; enforce governance for prompts, datasets, and versioning aligned with legal/compliance/audit.
Other
- Set the multi-year vision and portfolio roadmap for Document Intelligence, Intelligent Knowledge Retrieval, and AI Bots aligned to Guardian’s business strategy and digital transformation.
- Be the AI thought leader: identify methods, platforms, and operating models to scale GenAI/ML delivery; articulate build-vs-buy decisions and investment cases.
- Define north-star metrics (e.g., straight-through-processing %, cycle-time reduction, cost-to-serve, accuracy/quality, call deflection) and track value realization with executive stakeholders.
- Lead, mentor, and grow a high-performing team of data scientists; instill a culture of curiosity, rapid experimentation, and delivery excellence.
- Partner with Product, Engineering, Data, and business leaders to prioritize the backlog, integrate into workflows, and standardize reusable libraries, patterns, and reference architectures across the enterprise.
