Technical Lead, AI Engineering (Sales, Service & Supply Chain)

About the role 

Lucid is putting an AI layer over the systems that run our commercial and industrial operations, how we sell vehicles, how we service them, and how we move parts through a global supply chain. This role sets the technical direction for that work and is accountable for what reaches production. 

You will sit close to the business. Sales, Service and Supply Chain leaders will bring you problems, and you will decide what is worth building, design it, build the hard parts yourself, and lead a distributed engineering team to deliver the rest. This is a hands-on lead role, architecture and code, not a management track. 

What you'll do 

• Own the architecture for LLM-based and agentic systems across Sales, Service and Supply Chain: assistants over enterprise knowledge, agents that act against internal APIs, document and data extraction pipelines, and the platform pieces they share. 

• Deliver those systems into the channels our customers and employees already use: web and in-app chat, email and SMS. Own the conversational contract across them, including threading, context carry-over, handoff to a human, and tone appropriate to the channel. 

• Turn business problems into engineering plans. Work directly with stakeholders to define the outcome, then sequence the work so value lands early and often. 

• Build the parts that make AI safe to run in the enterprise: input and output guardrails, identity-aware access to tools and data, prompt governance, and clear trust boundaries between internal and external-facing surfaces. 

• Own the tool layer, including MCP servers and internal tool interfaces, so agents reach business systems through contracts that are versioned, permissioned and testable. 

• Define how quality is measured, including evaluation datasets, regression suites, tracing and production telemetry, and make it the default gate for prompt and model changes. 

• Own the unit economics. Set targets for cost and latency per conversation and per resolution, and hold them through model routing, context trimming, caching, batching and right-sized retrieval. 

• Integrate with systems of record: ERP, supplier portals, service and CRM platforms, logistics and planning systems. The interesting problems here are data, integration and correctness more than model choice. 

• Set standards the wider team builds on, such as shared services, reusable patterns and review expectations, and make sure they hold up outside your own code. 

• Lead technically across time zones, including close partnership with our engineering team in India. Mentor engineers, review designs, and be the person others bring their hardest problem to. 

• Be honest about where AI helps and where it does not. Part of the job is preventing expensive detours. 

What you'll bring 

• 10+ years building and operating production software, with 3+ years owning the technical direction of a team or a significant platform. 

• Strong Python and solid distributed systems fundamentals. Comfortable owning services end to end, including what happens after they ship. 

• Real production experience with LLM applications: retrieval-augmented generation, agent and tool-calling patterns, structured output, streaming, prompt design and iteration under load. 

• Strong hands-on experience with LLM engineering tooling such as Langfuse or LangSmith, used for tracing, prompt versioning, evaluation runs and production monitoring, not only local experiments. 

• Good working experience with MCP: building and consuming MCP servers, designing tool schemas, and handling authentication and permissions at the tool boundary. 

• Practical command of LLM cost optimization: model routing and tiering, context and prompt compression, prompt caching, batching, token budgeting, and knowing which lever to pull when spend moves the wrong way. 

• Experience delivering AI over customer and employee communication channels such as chat, email and SMS, and the integrations behind them. 

• Judgment on the model landscape, across hosted models, managed services and fine-tuning, enough to make a defensible choice on cost, latency and accuracy and move forward. 

• Strong AWS engineering (or equivalent): serverless and container workloads, event-driven design, IAM, networking basics, and a working sense of what things cost at scale. 

• Experience with the data layer behind AI systems, relational, document and vector, and with keeping retrieval accurate as content and users grow. 

• A concrete point of view on evaluating AI systems, and the discipline to build the harness before scaling the feature. 

• Experience working with non-technical stakeholders and translating between business intent and system design. 

Nice to have 

• Automotive, manufacturing, logistics or another operations-heavy domain background. 

• Exposure to SAP, supply chain execution, EDI, or field service and CRM platforms. 

• Hands-on work with messaging and contact center platforms such as Twilio, Amazon SES or Pinpoint, Salesforce Service Cloud, or similar. 

• Security-sensitive AI work: prompt injection defense, tenant or supplier data isolation, red-teaming. 

• Multi-agent orchestration, or experience running an evaluation program at scale. 

• Experience leading distributed teams across US and India time zones. 

How we work 

Small teams with direct ownership and a short path from a business problem to a running system. We prefer working software and measured results over decks, and we expect engineers at this level to have opinions and defend them with evidence.