Forward Deployed Engineer

About the Role

The AI CoE builds and scales AI solutions across CJ Express's business units. Forward Deployed Engineers (FDEs) are responsible for turning an available AI capability into a productionized solution that a business unit is actually running in production or facilitating to achieving this outcome with another specialized team. This individual will work directly with business unit stakeholders to scope real problems, build or adapt working AI solutions, and support their transition into production use, then train the business unit team to own and maintain the solution going forward, if applicable.

The FDE will consult, prototype, partially embed when a deployment requires sustained hands-on engineering, and hand off the solution with sufficient training so the business unit can operate and extend it independently. This is not a sales engineering role, as the responsibility extends beyond demonstrations. It is also not a full embedding role, as the individual will not be assigned indefinitely to a single business unit.

This role is part of building the AI CoE's forward deployment function from the ground up. The individual will help define how the AI CoE engages with business units and materialize AI transformation, rather than simply executing an established playbook.

Responsibilities

• Meet with business unit stakeholders to translate vague or loosely defined problems into a scoped, buildable AI solution.
• Build and ship working prototypes and production integrations, potentially including agents, RAG pipelines, data connectors, and internal tools, using agentic coding tools such as Claude Code to accelerate delivery and compliance with internal best practices.
• Partially embed with a business unit when development and deployment requires sustained hands-on engineering; otherwise operate under a consult-and-build model.
• Conduct training sessions and produce documentation so business unit teams can operate, troubleshoot, and iterate on the solution after handoff.
• Navigate CJ's data access, security, and PDPA requirements to bring a prototype to a deployment that is authorized to run in production.
• Contribute reusable patterns and components back to the AI CoE so that solutions built for one business unit can be applied elsewhere.
• Remain available after handoff to support scaling or resolve issues, with accountability for outcomes rather than delivery alone.

Additional responsibilities for Senior candidates:

• Lead multi-business-unit or higher-ambiguity engagements where the problem has not yet been scoped and the business unit has not yet identified how AI could be applied.
• Mentor mid-level FDEs on stakeholder management and deployment judgment, including decisions on whether to build, buy, or decline a request.
• Represent the CoE in cross-business-unit prioritization discussions.

Requirements

Must-have

• A minimum of 3 years of related experience for mid-level candidates, or 5 years for senior candidates.
• Strong proficiency in Python and working comfort with cloud infrastructure such as AWS, GCP, or Azure.
• Hands-on, working experience with agentic coding tools such as Claude Code or Cursor, rather than familiarity alone.
• Demonstrated ability to work directly with non-technical stakeholders, including scoping problems, managing expectations, and explaining trade-offs without technical jargon.
• Demonstrated comfort operating under ambiguity, as assignments will often begin as a business problem rather than a defined specification.

Preferred

• Experience building LLM-based applications, including agents, RAG, and tool-use or orchestration frameworks such as LangGraph or ADK.
• Experience with frontends and/or backends.
• Thai language proficiency is a must.

Bonus

• Experience with traditional AI and machine learning, including computer vision and recommendation systems or other deep learning modeling.
• English proficiency.
• Experience in retail, logistics, or FMCG domains.

Success Criteria

• Business unit teams are able to run AI solutions independently within an agreed handoff window, for example ninety days post-deployment, with the FDE serving as an escalation point rather than an operational bottleneck.
• Solutions are delivered as working production integrations rather than presentations or unmaintained prototypes.