Principal Engineer

Purpose of the role:

The role builds and owns the engineering harnesses on which AI-assisted delivery depends: the agent scaffolding, context and retrieval layers, evaluation and guardrail frameworks, and the telemetry that proves whether the model is working. It carries no line management accountability. Its authority derives from technical credibility, demonstrated results and delegated technical standards ownership, exercised across four technology stacks and every delivery team on the site.

This is a hands on technical role, strategic contributor. The postholder is expected to write production code, ship the tooling, and remain the deepest technical voice in the room on both the AI stack and the business domains it is being applied to.

The site operates four technology stacks, Java (core platform and services), .NET (back-office management platform), React (web) and Flutter (mobile),  across roughly 90 engineers and testers within a total organisation of c. 140. Engineering output is subject to licence conditions, safer-gambling technical requirements, data protection obligations and group information security standards.

The operating model is being redesigned around AI augmentation, with an expectation of materially higher throughput per engineer and a corresponding reduction in squad sizes over time. There is no validated sector reference implementation. The approach must be devised, instrumented, proven and adapted in operation, against a live estate carrying regulatory consequence.

Key accountabilities:

Each accountability is stated as outcome, authority and measure. Where accountability is genuinely shared, it is stated as shared.

1. AI engineering harnesses: Primary. Accountable for the design, build, operation and lifecycle of the internal harnesses that make AI-assisted engineering safe and repeatable at Rank: agent and assistant scaffolding, repository and domain context layers, prompt and workflow libraries, evaluation harnesses, regression and guardrail suites, and the integration of these into the developer toolchain and CI/CD pipeline. Decides the technical design and implementation of these systems. Measured by: harness availability and adoption, evaluation coverage of AI-assisted output, defect and rework rate on AI-assisted changes, time from harness request to production availability.

2. AI adoption in delivery: Primary for technical enablement, shared for realisation. Accountable for ensuring AI augmentation actually lands in the delivery teams: reference implementations, integration patterns per stack, developer enablement, and removal of the technical obstacles that stall adoption. Works through Heads of Engineering and Engineering Managers, who retain accountability for their teams’ delivery outcomes. Measured by: active adoption rate across squads, proportion of eligible workflows AI-assisted, measured throughput and quality change attributable to augmentation, benefit realised against the committed transformation case.

3. Domain and technical depth: Primary. Accountable for holding sufficient depth in Rank’s business domains — player journey, account and wallet, bonusing and promotions, game integration, payments, responsible gambling controls, regulatory reporting — to judge where AI assistance is safe, where it is not, and what a correct output looks like. Owns the domain context and knowledge artefacts on which AI-assisted work depends. Measured by: accuracy of domain context assets, quality of technical adjudication in design review, absence of AI-attributable domain or regulatory defects reaching production.

4. Technical standards and architecture practice: Shared with Architecture. Accountable for engineering standards governing AI-assisted development: code provenance and attribution, review obligations for generated code, licensing and IP hygiene, secure coding under assistance, and the definition of done for AI-assisted change. Sets these standards within the group reference architecture and recommends changes to it where augmentation requires them. Measured by: standards adoption, audit and assurance findings, security review outcomes on AI-assisted code paths.

5. Engineering telemetry and measurement: Shared. Accountable for the technical instrumentation that makes engineering performance observable, delivery, quality and flow telemetry, and the pipelines and analysis layers that turn it into decision-grade evidence for the Managing Director and group stakeholders. Determines the technical implementation; the measurement framework and its interpretation are agreed with the Managing Director and Heads of Engineering. Measured by: telemetry coverage across squads and stacks, data quality, timeliness and usefulness of insight delivered to leadership.

6. Technical evaluation and vendor assessment: Recommends. Accountable for the technical evaluation of AI platforms, models, tooling and third-party services proposed for engineering use, including capability benchmarking against Rank workloads, security and data-handling posture, cost model and exit risk. Recommends to the Managing Director, who holds selection and contracting authority. Measured by: quality and defensibility of evaluations, accuracy of realised outcome against predicted benefit and cost.

7. Technical leadership without authority: Primary. Accountable for raising the technical ceiling of the engineering organisation: mentoring senior and lead engineers, leading design and architecture review, setting the standard for engineering craft, and building the internal advocacy that carries the augmentation model past initial resistance. Holds no line authority over any engineer. Measured by: capability uplift in senior engineering population, quality of design decisions across squads, peer assessment.

8. Prototyping and technical risk reduction: Primary. Accountable for de-risking novel technical initiatives ahead of committed delivery through working prototypes and spikes, and for producing a clear, evidenced recommendation to proceed, adapt or stop. Measured by: proportion of major initiatives entering delivery with proven technical approach, avoided rework, cycle time from question to evidenced answer.

Nature and complexity of problems:

The role operates where no established method exists. Group strategy and regulatory constraint define what must be achieved; almost nothing defines how. Problems are typically novel, multi-variable, and resolved by adapting principle to circumstance under commercial and regulatory pressure.

Designing evaluation and guardrail frameworks for non-deterministic AI output in an estate where defects carry licence consequence, with no accepted industry standard to adopt.

Building harnesses that work coherently across four stacks with divergent tooling, release governance and lifecycle economics, without fragmenting into four separate solutions.

Distinguishing genuine productivity gain from displaced effort and measurement artefact, in a programme whose business case depends on the difference being real.

Encoding sufficient domain knowledge — bonusing logic, wallet and payment integrity, responsible gambling controls — into context layers that AI assistance can use correctly, where the tacit knowledge sits across dozens of engineers and is largely undocumented.

Securing adoption from experienced engineers whose roles are being materially changed by the technology being introduced, with no line authority and against reasonable professional scepticism.

Judging where AI assistance must be constrained or excluded on regulatory, security or IP grounds, and defending that judgement against commercial pressure for faster adoption.

Sustaining pace against a moving external technology frontier while keeping internal harnesses stable enough for teams to depend on.

Knowledge, skills and experience:

(a) Technical and professional depth

Deep, current, hands-on expertise in modern software engineering, with genuine production depth in at least two of Java, .NET, React and Flutter and working command of the remainder — sufficient to set standards and adjudicate technical trade-offs across the estate without deferring to others. Demonstrated depth in cloud-native architecture (AWS), microservices, event-driven systems, API design and CI/CD engineering.

Substantive practical experience with the AI engineering stack: LLM application patterns, agentic workflows and tool use, retrieval and context engineering, prompt and workflow design, model evaluation and benchmarking, and the failure modes and safety considerations specific to generative systems in production. Experience building developer-facing platforms and internal tooling used by other engineers.

(b) Domain and regulatory understanding

The ability to acquire and hold deep domain understanding rapidly. Existing knowledge of online gaming, betting, financial services or another regulated, high-transaction-volume consumer domain is strongly preferred, along with an understanding of the technical control environment that regulation imposes — auditability, data protection, provenance and evidential requirements.

(c) Influence and communication

Critical. Every outcome this role owns is delivered through engineers and leaders it does not manage, and defended to stakeholders who are not technical. Requires the ability to build credibility with sceptical senior engineers through demonstrated work, to explain technical trade-offs to commercial and group leadership without dilution, to hold a technical position under pressure, and to write clearly enough that decisions survive without the author in the room.

(d) Experience typically required

10+ years in software engineering, with a sustained record of hands-on delivery at senior or staff-plus level.

Demonstrated ownership of platform, tooling or developer-experience systems used across an engineering organisation of comparable scale.

Practical, shipped experience applying AI or machine learning to engineering workflows — not evaluation or advisory work alone.

Evidence of technical leadership without line authority across multiple teams.

Delivery accountability in a regulated, consumer-facing, high-availability environment (advantageous).

Bachelor’s or Master’s degree in Computer Science, Software Engineering, Engineering or a related field.

Key relationships:

Internal: Cape Town: Managing Director (line); Heads of Engineering; Architecture leadership; Quality Engineering; Product and Delivery leadership; Engineering Manager and Technical Leads; senior engineering population across all four stacks.

Internal: Group (UK): Chief Product and Technology Office; Director of Technology; Director of Architecture; Group Information Security; Group Data and Analytics.

External: AI platform and model vendors; developer tooling and cloud providers; third-party engineering service providers; security assurance and audit partners; the wider engineering community, where the role represents Rank technically.

Success in the first twelve months

The following describes what good looks like at the end of year one and forms the basis of the initial objectives.

A working, adopted AI engineering harness in production use across at least two stacks, with evaluation coverage and guardrails in place.

Domain context assets established for the highest-value areas of the estate, validated by the teams that own them.

Engineering telemetry instrumented sufficiently to evidence throughput and quality change with credibility at group level.

Published engineering standards for AI-assisted development, adopted across squads and accepted by Architecture and Information Security.

A measurable, defensible productivity change in the adopting teams, with the measurement method understood and trusted by leadership.

A visibly stronger senior engineering population, evidenced by the quality of design decisions being taken without escalation.