AI Workflow Engineer

About The Job:

• The AI Workflow Engineer is an integral part of a small expert client facing engagement team that will deliver business transformation through the AIVA AI orchestration platform. The AI Workflow Engineer is the builder at the heart of the AiVA engagement working with client sme’s, operational sme’s, the delivery principle and the AIVA tech-lead to automate and optimise client processes.
• During the client engagement, this role turns approved business process use-cases into working, governed AI skills inside the AiVA platform. The role will form part of the bootcamp team which will operate live, in the room, in front of client stakeholders, under time pressure, as well as delivering ongoing change throughout the engagement. The role is accountable for its robust, tested, deployable skills that deliver measurable business outcomes through the engagement lifecycle.

Overall aims:

• Design and implementation of governed AI workflows delivered against approved use cases which are defensible under client scrutiny. The workflows should maximise the re-use of existing skills where possible. Failing that the Workflow engineer will develop custom skills in line with best practice and submit for inclusion in the central catalogue
• Rapid iteration cycles held during the initial deployment phase: skills built, tested, and demonstrated within the compressed engagement window.
• Clear demonstrations of the platform capability, building awareness across sme’s and client staff to drive the effectiveness of the change programme.
• Develop skills where their performance and quality holds through initial deployment, sustained delivery, and expansion without rework.
• Working platform capability that earns client trust and unlocks expansion.
• Skills built to production quality: tested, governed, and maintainable.

Essential Job Functions:
Skill Design & Build

• Design and compose workflows on the orchestration canvas, wiring existing catalogue skills first and building new skills only where there is a genuine capability gap to avoid creating a large client catalogue
• Design prompts, workflows, tool integrations, and evaluation harnesses that turn each use case into a working, governed skill.
• Iterate rapidly during the initial deployment phase: build, demonstrate to client stakeholders, refine, redemonstrate within the compressed engagement window.
• Instrument skills for observability, evaluation, and continuous quality measurement across the engagement lifecycle.
• Design guardrails and safety patterns that hold under client testing and production load.
• Own skill quality end-to-end: functional correctness, evaluation coverage, integration behaviour, deployment readiness.
• Present any custom skills and workflows to the change board for design approval.

Engagement Execution

• Work as a fully-dedicated member of the multi-disciplinary delivery team through the engagement lifecycle until the engagement reaches sustained delivery.
• Partner with the Technical Lead on architecture decisions; own build execution within the agreed architectural pattern.
• Partner with the Business Analyst on requirements translation and with the Operations SME on domain understanding.
• Demonstrate working skills live to client stakeholders including senior operational and technical audiences.
• Coordinate with the Integration Engineer on data pipelines, upstream systems, and deployment infrastructure.
• Contribute technical input to scope and value conversations; flag build feasibility, timeline realism, and platform limits early.

Craft Development

• Develop deep AiVA platform fluency; contribute back to platform tooling and to the wider AI Workflow Engineer community through structured knowledge sharing.

Continuous Improvement & Capability

• Contribute reusable patterns, prompt templates, and evaluation frameworks to the platform capability library that compounds across engagements.
• Mentor newer AI Workflow Engineers as the function scales; contribute to capability development across the cohort.
• Ensure compliance with client requirements, data protection, security standards, and AiVA platform governance across all skills built.

Qualifications:

• Skill Building Craft - Designs, builds, tests, and deploys AI skills at pace with production quality. Comfortable in Skill Builder, Connector Builder, and AiVA Studio. Chooses the right prompt design, tool integration, and evaluation pattern for the use case.
• Rapid Iteration Under Pressure - Works live in the room during bootcamps and pilots; builds a working demonstration inside a compressed window. Comfortable rebuilding on the fly when client feedback shifts the requirement.
• Platform Fluency - Deep working knowledge of AiVA platform capabilities and limitations. Recognises what the platform does well and what it does not, and shapes technical proposals accordingly.
• Evaluation and Guardrails - Instruments skills for evaluation; designs test cases that stress the skill under realistic conditions. Builds guardrails that hold under testing and production load. Build in Human-In-The-Loop integrations as and when appropriate.
• Multi-Disciplinary Collaboration - Works effectively with Technical Lead, Business Analyst, Integration Engineer, and Operations SME. Comfortable translating between technical, business, and domain conversations without losing precision.
• Client Demonstration - Presents technical work credibly to senior client stakeholders including operational leaders, technical audiences, and non-technical sponsors. Comfortable demonstrating live under stakeholder scrutiny.
• Prompt Engineering - Designs prompts that produce reliable, evaluable, and governable behaviour from the underlying model. Understands prompt sensitivity, few-shot design, and structured output patterns.
• Continuous Learning - Reflects on skills built; adapts patterns based on what works and what doesn't; contributes reusable components back to the platform library and community.

Personal Specifications/Attributes

• Strong technical builder mindset with the discipline to ship rather than to perfect.
• Comfortable operating under sustained pressure in client-facing rooms; builds live, demonstrates live, iterates in real time.
• Deep interest in AI-enabled automation as a craft; keeps up with the platform, the model landscape, and the prompt engineering community.
• Process instinct: notices where a workflow breaks in practice, not just where it's defined to break. Thinks about exception handling, human touch-points, and handoffs as central to skill design rather than as edge cases.
• Collaborates well across technical and non-technical disciplines; comfortable translating between architecture, requirements, and domain conversations.
• Resilient under change; iterates gracefully when client feedback shifts a requirement mid-build.
• Continuous learner; adapts patterns based on what works and what does not.
• Quality-first orientation; treats evaluation and guardrails as first-class deliverables, not afterthoughts.
• Comfortable working across geographies (US, UK, India) and across sectors including legal and professional services.

Key Requirements

• 3+ years of software engineering, applied AI or machine learning, or automation platform development experience.
• Demonstrated experience building production-quality AI applications, including prompt engineering, tool integration, and evaluation.
• Familiarity with modern LLM APIs, agent frameworks, and orchestration platforms; willingness to develop deep AiVA platform fluency rapidly.
• Experience working in client-facing or client-adjacent delivery roles preferred; comfortable presenting technical work to senior audiences.
• Bachelor’s degree in Computer Science, Engineering, or a related technical field, or equivalent experience.
• Strong written and verbal communication skills; can explain technical decisions to non-technical audiences.
• Comfortable working across geographies and time zones (US, UK, EMEA clients).
• Operational or process-oriented background is highly desirable: experience working in, alongside, or embedded with an operational team with exposure to how workflows actually run, including where they break.
• Process mindset: designs skills that anticipate exceptions, handoffs, and human-in-the-loop patterns rather than just the happy path. Comfortable modelling a workflow as it is actually performed, including the parts that don't appear in the spec.