AI Engineer

Eloa Labs; Lisboa; On-site; EUR 60,000 - 90,000; Full time

Job summary

Eloa Labs is seeking an AI Engineer to build and improve AI pipelines, turning requirements into tested software and delivery-ready code. You will develop both the factory and adjacent projects, using AI harnesses to implement changes within existing codebases and architectures.

You will ensure correct behavior, security, and performance, while guiding tools and collaborating with stakeholders to meet requirements and timelines. Acceptance criteria and regression checks are built into workflows.

Qualifications

• Strong CS fundamentals in algorithms, data structures, systems design, testing, and debugging.
• Experience delivering and maintaining production software, including APIs and CI/CD.
• Proficiency in at least one language (Go, Rust, TS, Java, Python) and ability to learn others.
• Experience with AI coding assistants or harnesses to deliver software.
• Knowledge of LLM APIs, prompts, context, and multi-step agent workflows.
• Ability to diagnose defects across application code, integrations, and AI workflows.
• Experience navigating unfamiliar codebases with respect to architecture, interfaces, and behavior.
• Understanding of databases and performance considerations.

Responsibilities

• Design, build, operate, and improve AI pipelines that generate, transform, validate, test, and support the deployment of software.
• Read, review, debug, and directly modify generated code to correct defects and improve architecture, security, and performance.
• Reproduce failures, investigate root causes, and add regression tests for pipeline improvements.
• Implement features and refactor projects using AI harnesses outside the factory, respecting business rules and interfaces.
• Guide AI tools with clear tasks and project context, review outputs, and assess risks before delivery.
• Integrate LLM APIs, agents, developer tools, repositories, CI/CD, databases, and internal services.

Skills

CS fundamentals

Production software

Programming languages

AI tooling

LLM APIs

Tools

CI/CD

Git

Databases

Job description

We are looking for an AI Engineer to build and improve AI development pipelines and deliver reliable software using AI tools. The role combines development of an AI software factory, hands-on engineering of the software it produces, and development of existing brownfield projects using AI coding harnesses.

The software factory uses AI pipelines to turn requirements into implemented, tested, and deployable software. You will develop and maintain these pipelines and work directly on their output when generated code contains defects, misses requirements, or needs further engineering. You will also develop projects outside the factory, using AI coding harnesses to implement changes within existing codebases and architectures.

This role requires strong software engineering fundamentals, proficiency in at least one programming language, and the ability to become productive in others. You will take responsibility for understanding, validating, and maintaining the software you deliver. Depending on project needs, assignments may also include working directly with customers in a forward deployed capacity.

Key Responsibilities

• Design, build, operate, and improve AI pipelines that generate, transform, validate, test, and support the deployment of software.
• Read, review, debug, and directly modify generated code to correct defects, close requirement gaps, and improve architecture, security, performance, and maintainability.
• Reproduce failures, investigate root causes, and add regression tests. Feed recurring issues back into pipeline logic, instructions, context, and evaluations.
• Implement features, fix bugs, and refactor existing projects using AI coding harnesses outside the software factory. Understand business rules, architecture, dependencies, and interfaces before making changes.
• Guide AI tools with clear tasks and relevant project context, review their output, and verify behavior, compatibility, and regression risks before delivery.
• Integrate LLM APIs, agents, developer tools, repositories, CI/CD systems, databases, and internal services.
• Build automated evaluations, tests, quality gates, and appropriate human review into AI workflows. Improve observability, failure recovery, latency, and cost efficiency.
• Apply secure coding practices, access controls, sandboxing, and secrets protection when AI tools generate or execute code or access systems.
• Collaborate with architects, developers, and project stakeholders to clarify requirements, agree acceptance criteria, integrate changes, and resolve issues through release and stabilization.
• Evaluate new AI tools and engineering practices for practical adoption. Where required by an assignment, work with customer teams on requirements, solution integration, deployment, documentation, and handover.

Required Skills and Experience

• Strong computer science and software engineering fundamentals, including algorithms, data structures, systems design, testing, and debugging.
• Experience delivering and maintaining production software, including APIs, backend services, databases, version control, code review, and CI/CD.
• Strong proficiency in at least one programming language and the ability to learn and work across others. Projects may use Go, Rust, TypeScript, Java, Python, or other languages.
• Practical experience using AI coding assistants, agents, or harnesses to deliver software, with the judgment to inspect, challenge, and correct their output.
• Working knowledge of LLM APIs, prompting and instructions, context management, structured outputs, tool calling, and multi-step agent workflows.
• Ability to diagnose defects across application code, integrations, and AI workflows, distinguishing software defects from pipeline and environment failures.
• Experience navigating unfamiliar codebases and making changes that respect established architecture, interfaces, business rules, and behavior.
• Understanding of relational and non-relational databases, data modeling, queries, indexing, and performance considerations.
• Knowledge of automated testing, static analysis, observability, secure coding, data access, secrets handling, and the limits of AI-generated output.
• Clear communication, ownership of delivery, and the ability to clarify ambiguous requirements, explain technical tradeoffs, and adapt to new tools.

Preferred Qualifications

• Experience building AI development pipelines, software factories, developer platforms, or engineering automation.
• Experience designing AI evaluations and regression datasets, or measuring workflow quality, reliability, latency, and cost.
• Familiarity with containers, isolated execution environments, cloud platforms, infrastructure-as-code, and deployment automation.
• Experience with retrieval, embeddings, vector databases, code indexing, or orchestration frameworks such as LangChain, LlamaIndex, Semantic Kernel, or similar.
• Customer-facing or forward deployed experience, including discovery, solution integration, technical demonstrations, and operational handover, is beneficial but not required.

Ideal Candidate Profile

The ideal candidate is a hands-on software engineer who is comfortable building AI pipelines, correcting the software they generate, and developing existing applications. They can understand unfamiliar systems, work across languages, and use AI tools with sound engineering judgment. They take responsibility for delivered code, investigate failures thoroughly, and turn recurring problems into improvements to tools and workflows. They combine adaptability and clear communication with attention to detail, security, and maintainability.

Success in This Role Looks Like

• AI development pipelines produce reliable, maintainable software with fewer recurring defects and less avoidable rework.
• Delivered software meets agreed requirements and quality standards, supported by automated checks, tests, reviews, and appropriate controls.
• Defects are resolved with regression coverage, and recurring problems lead to relevant pipeline improvements.
• Brownfield projects progress through verified features, fixes, and maintainable changes supported by effective use of AI harnesses.
• Improvements are demonstrated through relevant measures such as task success rate, escaped defects, delivery time, and cost per successful task.