AI Engineer / Forward Deployed Engineer

oryxsearch.io; Abu Dhabi; On-site; AED 240,000 - 480,000; Full time

Job summary

oryxsearch.io is seeking an AI Engineer / Forward Deployed Engineer to join a small team of senior engineers working inside government entities to redesign workflows and deploy agentic AI in production.

You will own domain rewiring, adoption and realised value, collaborating with AI, product, platform architecture and evaluation specialists. This is an individual contributor role with potential to lead a rewiring pod as the programme grows.

Qualifications

• Shipped software end-to-end in production.
• Experience in embedded/stakeholder environments.
• Experience with agentic AI and LLM systems.
• Ability to interview users, define requirements and trim scope without a PM.

Responsibilities

• Own domain rewiring, adoption and realised business value.
• Work with AI engineering, product engineering, platform architecture and evaluation teams.
• Potential to lead a rewiring pod as the programme grows.
• Define users, workflow, scope, success metrics and value case.

Skills

Agentic AI

LLM

Orchestration

Retrieval

Evaluation

Prompt engineering

Azure

Arabic & English bilingual

Tools

Azure

Job description

AI Engineer / Forward Deployed Engineer - AI & Agentic Systems

About the Role

Domain Rewiring means redesigning the operational domain itself — workflows, decision points, data flows and human roles — so that agentic AI becomes embedded into how work actually gets done.

The Forward Deployed Engineering team works directly inside government entities to identify, build and deploy these systems, ensuring the underlying AI platform is shaped by real operational requirements rather than developed in isolation.

Why Join

• Work on complex applied AI problems. Re-engineer live government workflows and deploy agentic AI where operational decisions are actually made.
• Own outcomes, not just engineering. Define the users, workflow, success metrics and scope, then take the solution through to production and adoption.
• Small, senior teams. Work in compact squads of experienced engineers with significant autonomy and direct access to senior government stakeholders.
• Build the practice. Help establish the playbook for evaluation, AI-native software development and forward deployment, before helping grow the team around it.
• UAE-based opportunity. Relocation assistance is available for successful candidates and their families.

The Role

We are hiring a AI Engineer / Forward Deployed Engineer to join a small team of senior engineers responsible for taking complex government domains from discovery through to production AI systems.

This is an individual contributor role first.

You will personally own some of the organisation’s most challenging domain rewirings, working directly with users and senior stakeholders to understand how an operation works today and redesign how it should operate in an AI-native environment.

There is no traditional business analyst, product manager or architect sitting between you and the problem. You define the users, workflow, scope, success metric and value case — and then build.

You will own domain rewiring, adoption and realised business value, while partnering with specialists across AI engineering, product engineering, platform architecture and evaluation.

As the programme grows, you may also lead a rewiring pod, hire engineers or own part of the domain portfolio. However, leadership sits alongside hands-on delivery rather than replacing it.

What You Own

• Embed and define. Work directly inside government entities, understand how operations actually run and define the requirements yourself. Establish the metric, capture the baseline and agree it with the business owner before building.
• Design and ship. Architect agentic workflows spanning orchestration, retrieval, tools, agents, evaluation and human-in-the-loop processes — then write the production code.
• Navigate enterprise constraints. Take deployments through security, privacy, tenant isolation, governance and information-security reviews, creating reusable patterns for future deployments.
• Own adoption. Redesign the human process alongside the software. Success means changing how decisions are made, reducing cycle times, lowering costs or improving measurable outcomes.
• Prove value. Instrument products so their impact can be measured against an agreed baseline. Recommend scaling, redesigning or stopping initiatives based on evidence.
• Make the next deployment faster. Convert lessons from individual rewirings into reusable platform capabilities, engineering patterns and playbooks.

First Three Months

Indicative outcomes include:

• Delivering one lighthouse domain rewiring end-to-end within a major government entity, reaching production and demonstrating measurable improvement.
• Establishing an agreed value case and baseline with the relevant business owner and delivering the first value report.
• Contributing reusable capabilities to the underlying sovereign AI platform based on requirements discovered through real-world deployment.

Who You Are

Strong production engineer.

You have personally shipped software end-to-end. Years of experience matter less than what you have successfully put into production.

Field-hardened.

You thrive in ambiguous, embedded and stakeholder-facing environments where the problem has not yet been clearly defined.

Hands-on with agentic AI.

You have practical depth across LLM and agent systems, including orchestration, retrieval, evaluation, agent design and prompt engineering. AI coding tools are part of your normal development workflow.

Product owner by instinct.

You are comfortable interviewing users, defining requirements, prioritising problems and cutting scope without waiting for a Product Manager.

Business literate.

You can discuss cost, cycle time, productivity, headcount and ROI with senior executives without hiding behind technical terminology.

Enterprise literate.

You understand identity, networking, data governance, security and approval processes required to move software into production within complex organisations.

Value disciplined.

You are willing to stop, redesign or descope your own work when the evidence shows it will not create sufficient value.

Strongly Preferred

• Forward Deployed Engineering, Solutions Engineering or technical delivery experience.
• Experience building or leading an FDE capability.
• Delivery experience within government, financial services or another regulated/high-assurance environment.
• Founder, founding engineer or zero-to-one product experience.
• Hands-on cloud and platform engineering experience, particularly Azure.
• Arabic and English capability, or experience operating effectively within bilingual environments.

Evidence We Will Ask For

We will want to understand:

• Two products you personally built that reached real users in production, including your contribution, scale and measurable outcome.
• For at least one product, the baseline before deployment, measured result afterwards, how the impact was measured and who validated it.
• One product you stopped, redesigned or significantly descoped because the expected value did not materialise, including the evidence behind that decision.

Projected benefits in a business case are not considered realised impact.

How Performance Is Measured

Performance is based on three primary outcomes:

1. Speed: Time from problem definition to a rewired workflow being used in production.
2. Value: Measurable impact against an agreed baseline, validated by the relevant business owner.
3. Judgement: Rewirings that successfully pass the value checkpoint — and initiatives correctly stopped early when the evidence does not support further investment.

Performance is not measured by lines of code, number of pilots or size of team.

What This Role Is Not

Not a management role disguised as engineering.

You may lead engineers or a portfolio, but you remain personally responsible for delivering rewirings.

Not advisory.

Recommendations that end with a presentation or strategy document do not count. You build and ship.

Not a pilot factory.

A portfolio of proofs of concept that never reaches production or creates measurable value is not considered successful.