AI Engineer Forward Deployed

Job Description

Mission

Join the Engineering team as a Forward Deployed AI Engineer a hybrid role combining the skills of an engineer, solutions architect, and consultant. This position is designed to embed directly with internal operational teams (beginning with Managed Services) to develop, implement, customize, and troubleshoot AI models in real-world production environments. You will serve as the critical bridge between our product team and internal operational departments, translating cutting-edge AI capabilities into practical, measurable outcomes for the business. The ideal candidate thrives in ambiguous, fast-moving environments and is equally comfortable writing production code, advising stakeholders, and redesigning workflows around AI-first thinking.

Position Overview

Embedded operational deployment: Act as a resident AI expert within Individual Departments/Business Units (e.g.Managed Services) and other internal teams, understanding their workflows end-to-end and identifying where AI can drive efficiency, accuracy, and scale.

Hybrid engineer-consultant model: Function as engineer, solutions architect, and internal consultant designing solutions, building them, and guiding teams through adoption and change management.

LLM application development: Design and build AI-powered application features using LLM APIs, tool-calling patterns, and modern coding tools.

Agentic workflow ownership: Create agent loops that can select tools, execute actions, summarize results, and produce traceable user responses.

AI chat interface focus: Develop chat-based analytical experiences that connect user questions to backend tools, data services, semantic models, and visualization outputs.

Prompt and cost optimization: Improve prompt quality, reduce token usage, manage context windows, and optimize model/API cost without degrading output quality.

Modern engineering productivity: Use advanced AI coding tools such as Cursor, Claude, Codex, or similar tools to accelerate development while maintaining code quality and review discipline.

Key Responsibilities

Embedded Team Partnership & Operational AI Enablement

• Embed directly with internal departments to understand day-to-day workflows, pain points, and operational bottlenecks where AI can have the highest impact.
• Act as the on-the-ground AI expert participating in team standups, process reviews, and strategic planning sessions to continuously surface AI opportunities.
• Translate operational needs into AI solution requirements, bridging communication between the product engineering team and internal stakeholders.
• Lead the end-to-end implementation of AI solutions within operational contexts: from scoping and design through build, testing, deployment, and iteration.
• Provide hands-on troubleshooting and support for AI models running in production within internal team environments, ensuring reliability and performance.
• Drive change management and adoption by training internal team members on new AI tools, workflows, and best practices.
• Document operational AI use cases, implementation patterns, and lessons learned to inform the product roadmap and support scaling to additional teams.

LLM Application and Agent Development

• Design, build, and maintain LLM-powered features for enterprise data applications, including natural-language analytics and AI-assisted workflows.
• Implement agent loops that support multi-step reasoning, tool-calling, retry handling, tool-result summarization, and final response generation.
• Define and maintain tool schemas for LLM tool-calling, including tool names, descriptions, required inputs, output contracts, and safe execution boundaries.
• Build orchestration logic that maps LLM tool requests to backend functions, executes the tools, handles errors, and feeds summarized results back into the conversation.
• Create traceable AI experiences where users can inspect tool steps, generated SQL, data outputs, chart recommendations, and final explanations.

Prompt Engineering, Model Usage, and Optimization

• Develop domain-aware system prompts, instruction templates, and response formats tailored to Managed Services workflows and broader pharmaceutical data analytics use cases.
• Optimize prompts for reliability, concise responses, controlled formatting, and consistent behavior across Quick Mode, Agentic Mode, and AI Chat experiences.
• Understand LLM context windows, token budgeting, tool result truncation, conversation memory, and prompt injection risks.
• Monitor and improve model performance through test cases, prompt evaluation, failure analysis, and iterative tuning.
• Apply token and cost optimization techniques such as compact tool results, selective context inclusion, response constraints, and model selection tradeoffs.

AI Chat Interfaces and User Experience

• Build production-quality chat interfaces that support user questions, streaming or step-based responses, history, reruns, contextual suggestions, and tool result display.
• Collaborate with product and operational teams to make AI outputs understandable, actionable, and trustworthy for business and technical users.
• Integrate chart recommendations, result tables, SQL expanders, and execution traces into the AI user experience.
• Design graceful error handling for model timeouts, malformed tool calls, invalid JSON responses, failed SQL, expired sessions, and partial agent results.
• Partner with SRE and security teams to define safe deployment, monitoring, logging, and operational support patterns for AI workloads.

Cloud Deployment and Engineering Practices

• Design and support AI application deployment patterns in AWS, including containerized services, API-based workloads, and secure integration with enterprise identity and data systems.
• Implement backend services and application modules using Python and modern API patterns.
• Use Git-based development, code reviews, automated checks, and documentation to support production-quality releases.
• Work with DevOps/SRE teams on environment configuration, secrets handling, observability, and runtime monitoring.
• Contribute to reusable AI engineering standards for prompts, tools, evaluation, logging, and deployment.