Principal Software Engineer / Architect – AI – Legal FDE

Responsibilities: Define and drive the long-term technical vision and roadmaps for FDE aligned to product goals Lead selection and adoption of technologies, building consensus across teams Identify opportunities in agentic systems, interoperability, and new model patterns, communicating trade-offs to leadership Contribute to cross-functional engineering strategy, standards, and patents where appropriate Lead complex AI initiatives end-to-end with hands-on design, prototyping, and coding Architect scalable, secure AI systems that integrate with client environments and legal tech ecosystems Design RAG pipelines, multi-agent orchestration, and agentic workflows using CoCounsel and other capabilities Build reusable AI components and templates; set standards for MLOps/LLMOps, IaC, and CI/CD Identify design weaknesses and drive remedies Set strategy for AI monitoring, evaluation, and production reliability Own LLM evaluation, benchmarks, and red-teaming; ensure compliance with ethics, data protection, and security Drive post-launch operational excellence and lifecycle metrics Act as senior technical voice in discovery, scope use cases, and shape architectures for high-value engagements Present complex concepts to engineers and executives, negotiating solutions between stakeholders Mentor engineers, review architecture, and strengthen AI-native culture Engage with program managers and budgets for large programs; represent Thomson Reuters externally

Key requirements: 12+ years software engineering with deep system design and distributed architectures Hands-on AI/ML and LLM production experience Expertise with orchestration frameworks (LangChain, LangGraph, PydanticAI, AutoGen) and RAG architectures with vector databases Experience with Model Context Protocol (MCP) and Agent-to-Agent (A2A) patterns Expert Python (FastAPI or equivalent) with RESTful, async, and event-driven design Cloud platform experience (AWS, Azure, or GCP) and Infrastructure as Code (Terraform, CDK, or equivalent) Production observability tools (Datadog, Grafana, Prometheus) and AI evaluation platforms (LangSmith, Langfuse, Braintrust, Weights & Biases) Secure API development (OAuth 2.0, OpenID Connect, JWT) Experience with AI-assisted development tools (Claude Code, Cursor, GitHub Copilot) and AI-native practices Strong influence across teams, executive communication, and ability to thrive in ambiguity Recognized technical leader and role model Strategic communication Mentorship and coaching Cross-functional collaboration LangChain LangGraph PydanticAI