Enterprise AI Architect

Bonapolia; Warszawa; On-site; PLN 250,000 - 450,000; Full time

Job summary

Bonapolia is seeking an Enterprise AI Architect to define the technical shape of the AI capability on an international engagement. You will work directly with the client's CTO and senior stakeholders, select platforms and approaches, and defend those decisions while staying hands-on to own code quality and user outcomes.

You will also design and build reusable AI skills, establish standards, drive presales activities, and mentor two engineers to deliver production-ready AI solutions with

Qualifications

• Hands-on experience with modern AI tools and platforms
• Real LLM-backed systems shipped to real users
• Working command of agentic patterns: tool use / function calling, MCP, prompt engineering, context engineering
• Sound judgement about applicability of techniques and when to use conventional software
• Experience evaluating non-deterministic systems with measurable criteria

Responsibilities

• Shape the work with senior stakeholders to establish objectives and outcomes
• Design, build, test and maintain reusable AI skills and components
• Write and review production-quality code and manage CI/CD
• Develop and maintain the evaluation layer with acceptance criteria and regression checks
• Provide technical leadership to two engineers and drive platform-level decisions

Job description

Enterprise AI Architect

Format: project for an international client; a three-engineer team (Architect + 2 Lead AI Solutions Engineers). Technical lead of the AI capability.

About the Role

We are looking for an Enterprise AI Architect to define the technical shape of the entire AI capability on the engagement: standards, shared components, and the evaluation and release strategy. You will work directly with the client's leadership (CTO, enterprise architects, senior business stakeholders), select platforms and approaches, and defend those decisions in front of both technical and business audiences. You stay hands-on throughout: you write and review code and personally own the quality of what reaches users.

What You Will Do

Shape the work

• Work with senior business stakeholders and domain experts to establish objectives, constraints, and expected outcomes.

• Identify which processes are worth automating and which are not, and say so plainly, including when the answer is no.

• Select the right tools, platforms, and approaches for each business problem and justify the choice.

• Shape and present solution proposals: scope, approach, effort, expected impact. Presales is part of the job.

• Own outcomes with a product mindset: adoption, measurable time saved, and the reliability of what real users experience.

Build it

• Design, build, test, and maintain reusable AI skills: instructions, supporting scripts, validation logic, reference data, shared components.

• Write and review production-quality code, owning it through review, CI, and release.

• Build the evaluation layer: golden cases, acceptance criteria, regression checks.

• Debug agent behaviour: establish why a workflow produced the wrong result and fix the root cause.

• Integrate with the client's platforms and data sources.

Drive it forward

• Run pilots with real consultants, collect structured feedback, and ship improved versions.

• Set and enforce standards for how skills are authored, reviewed, versioned, and released.

• Track where AI tooling is heading and bring the relevant parts into the engagement before the client asks.

• Provide technical leadership to two engineers: review, direction, unblocking.

What We Expect

Architecture and engineering

• Proven architectural ownership of real systems: module boundaries, reuse, versioning, integration, non-functional constraints.

• Still hands-on: you build, review, test, and maintain production-quality solutions personally. This is not a role for pure managers or "PowerPoint architects".

• Strong engineering foundation in at least one modern stack. Our delivery stack is TypeScript / Node.js (NestJS, Angular); fluency is an advantage, not an entry condition. A strong architect from a Python, Java, or .NET background who is fluent in modern AI tooling is a great fit.

• Standard professional practice: version control, code review, automated testing, CI/CD.

AI engineering

• Hands-on experience with modern AI tools and platforms, and a working map of the current landscape rather than familiarity with a single vendor.

• Real LLM-backed systems shipped to real users, not prototypes.

• Working command of agentic patterns: tool use / function calling, MCP, prompt engineering, context engineering, structured instruction design, orchestration.

• Sound judgement about applicability: which business problems fit which technique, and where the honest answer is conventional software.

• Experience evaluating and improving non-deterministic systems: test sets, measurable acceptance criteria, iteration driven by results rather than intuition.

Architect-specific

• Make and justify platform-level decisions: agent harnesses, integration architecture, build-versus-adopt.

• Define and enforce what a well-formed skill looks like, what belongs in shared infrastructure, and what gets rejected in review.

• Data handling, security, and compliance judgement when working with a client's sensitive HR and compensation data.

• Credible in the room with a client's CTO, enterprise architects, and senior business leadership.

Consulting and communication

• Able to explain a technical approach and its trade-offs to a non-technical audience and be understood.

• Presales capability: shaping solutions, scoping, estimating, and presenting to decision-makers.

• Able to turn ambiguous business requirements and tacit domain knowledge into a clear design and an executable workflow.

• Comfortable operating without a specification and asking the questions that close the gap.

• English at C1: you run client conversations unaided.

• Sustained working-hours overlap with the client (US / Western Europe).

Nice to Have

• Hands-on experience with Claude Code, particularly authoring Claude Code Skills, plugins, or MCP servers.

• Experience with LLM evaluation tooling (promptfoo, LangSmith, custom harnesses).

• Consulting or professional services background.

• Domain exposure to HR technology, compensation & benefits, payroll, or workforce analytics.

• Experience building internal developer tooling or standing up a platform capability.

• Public artefacts: open-source agent tooling, published skills or plugins, conference talks, technical writing on agentic engineering.

  • Model: B2B contract with Bonapolia - one contract, one monthly invoice

  • Rate: negotiated per candidate seniority and client budget

  • Payment: monthly invoice, paid on the same schedule the client pays us (usually 30 days)

  • Contract written under your country's law

  • We handle the client-side paperwork: MSA, NDA, invoicing

  About Bonapolia

  Bonapolia is a family-owned IT staff augmentation team based in Lisbon. We build engineering teams for product companies and IT consultancies across Europe and the US.

  More at bonapolia.comReady to Join?We look forward to receiving your application and welcoming you to our team!