Staff Software Engineer/Data Scientist

We’re determined to make a difference and are proud to be an insurance company that goes well beyond coverages and policies. Working here means having every opportunity to achieve your goals – and to help others accomplish theirs, too. Join our team as we help shape the future.

Key Responsibilities

• Design & Deliver AI Solutions: Build statistical, ML, and generative/agentic AI solutions spanning RAG pipelines, chat/assistants, classification, forecasting, and recommendation systems using a fit‑for‑purpose toolkit from traditional predictive modeling to agentic workflows.
• Regulatory Intelligence & Filing Automation: Design and deploy GenAI capabilities to automate regulatory filing support, including DOI objection response generation and ingestion of legacy filings into searchable knowledge bases. Partner with Legal and Compliance to ensure outputs meet evolving standards and enable direct API integrations with regulatory bodies.
• Knowledge Base Engineering for Strategic Domains: Engineer and maintain domain‑specific knowledge bases (regulatory intelligence, competitive insights, customer sentiment) to power generative applications across underwriting, pricing, and service.
• Domain & Compliance Integration: Develop deep understanding of The Hartford’s business structures, processes, and data sources. Embed domain taxonomies, regulatory constraints, access controls, and security into the solution design. Ensure adherence to responsible AI practices—fairness, bias mitigation, transparency, and observability with compliance‑by‑design.
• Stakeholder Collaboration: Partner with leaders and SMEs across Product, Operations, Claims, Underwriting, and Risk to align initiatives to business goals. Define success criteria that balance accuracy, reusability, cost, and performance. Translate model behavior into actionable strategies with clear ROI.
• End‑to‑End Solution Development: Own the AI lifecycle from problem framing through deployment: data prep, modeling, evaluation, CI/CD, orchestration, observability, safety filters/guardrails, and rollback plans. Collaborate closely with AI engineers, data engineers, platform, security, and IT for seamless integration.
• Unstructured Data & Retrieval Design: Prepare multi‑format content (PDF, Office, HTML, images, audio) with normalization, metadata/lineage management, and PII detection/redaction. Design retrieval strategies (chunking, embeddings, hybrid search) and tune for cost, latency, and domain fit; leverage rerankers where appropriate.
• Prompt & Agent Design: Author robust system prompts, few‑shot patterns, and structured outputs (e.g., JSON schemas). Define safe tool‑use policies and function/structured calling for reliable agent behavior.
• Evaluation & Monitoring: Define metrics across use cases for classification, information retrieval, RAG/chat, forecasting, plus customer/ops KPIs. Build gold/synthetic test sets, support A/B testing, and monitor drift. Provide economic, qualitative, and statistical analysis supporting thresholds and decisions.
• Synthetic Data Generation & Augmentation: Develop and validate synthetic data pipelines to alleviate sparsity and accelerate convergence, especially for low‑frequency perils and emerging segments, while preserving privacy and distributional fidelity.
• Customer Experience Optimization: Apply GenAI to elevate self‑service, virtual assistants, and inspection automation, driving personalization, speed, and operational efficiency.
• Architectural Collaboration & MLOps Integration: Partner with enterprise architects and platform teams to ensure scalable, secure deployments via unified systems. Standardize experiment tracking, registries, evaluation gates, and CI/CD patterns across clouds and services.
• Innovation & Continuous Learning: Identify and pilot emerging methods (OCR, rerankers, PEFT/LoRA, distillation). Build reusable accelerators (chunking templates, prompt registries, evaluation harnesses). Stay current on AI/ML, LLMOps, NLP, RAG, and responsible AI.

Required Skills & Experience:

• Experience Range – 6 to 9 Years
• Experience in statistical modeling and machine learning using Python, including extensive use of pandas, NumPy, scikit-learn, and strong SQL for data exploration, feature development, and knowledge preparation; familiarity with PyTorch and/or TensorFlow preferred.
• Experience across the end-to-end modeling lifecycle, including problem framing and requirements gathering, experiment design, offline evaluation, and ongoing production validation and monitoring.
• Solid understanding and practical application of core machine learning methods, with 3+ years of experience applying deep learning architectures in real-world use cases.
• Experience designing and operationalizing model evaluation and monitoring approaches, including test set creation (gold and/or synthetic), metric definition and tracking (e.g., classification, forecasting, ranking/IR, and business KPIs), and supporting A/B testing, drift detection, and performance regression monitoring.
• Experience working with unstructured data, including document parsing and OCR fundamentals, text normalization, metadata and lineage awareness, and PII detection or redaction considerations.
• Experience using Git and Unix-based development environments, with experience building reproducible notebooks or pipelines and ensuring repeatable analytical workflows; 3+ years of exposure to basic container and cloud fundamentals supporting deployment workflows
• Experience communicating modeling decisions, design tradeoffs, evaluation results, and risks to both technical and non-technical audiences, and translating analytical outcomes into measurable business impact.
• Experience working with cloud-based AI platforms such as Google Vertex AI, AWS SageMaker or Bedrock, or Azure AI Services, supporting experimentation, model training, and deployment.
• Experience deploying models and integrating scoring logic into production systems, including operation within complex enterprise or packaged application environments (e.g., Duck Creek, Ratabase).
• Experience with NLP and Generative AI capabilities, including embeddings, retrieval strategies (dense and hybrid), chunking approaches, prompt engineering, structured outputs, and contributing to Retrieval-Augmented Generation (RAG) solutions and evaluations.
• Experience or exposure to advanced GenAI applications and extensions, such as agent or tool-use concepts, domain-specific knowledge graph integration, synthetic data generation, sentiment modeling, and GenAI use cases in filing or compliance contexts.
• Experience working within enterprise AI governance expectations, including aligning model development with compliance, privacy, documentation, and ethical standards.

Nice to Have

• RAG Expertise: Handson with vector databases and search (e.g., Vertex AI RAG Engine, OpenSearch, pgvector/Postgres), ANN indexing (HNSW), rerankers (crossencoders), and evaluation frameworks (RAGAS, TruLens, DeepEval).
• Document AI Tooling: PyMuPDF/pdfplumber, Apache Tika; OCR (Tesseract); layoutaware models (LayoutLM); table extraction (Camelot/Tabula).
• Embedding Model Selection: Experience comparing OpenAI/Cohere/Voyage vs. opensource (bge/e5/gte) for domain corpora; understanding dimension/quality/cost/latency tradeoffs and multilingual needs.
• Orchestration Frameworks: Familiarity with LangChain, LangGraph, or LlamaIndex; structured tool/function calling and guardrails for AI agents.
• CloudNative ML: Handson with Vertex AI, SageMaker, or Azure ML; experiment tracking (MLflow/W&B), registries, and CI evaluation gates.
• Responsible AI & Safety: Bias/fairness testing, hallucination mitigation, grounding checks, safety filters; basic model risk documentation.

• Broader Modalities (Nice to Have): Timeseries forecasting, recommenders, anomaly/fraud detection, speech/vision/multimodal.
• Fine tuning LLMs and Diffusion models using PEFT/LoRA, experience with distillation