Full-Stack AI Engineer
Mavila Consulting Remote Remoto Publicada em 19 de jun. de 2026
Senior
A descrição completa está disponível na vaga original.
Responsabilidades
- Design and build systems that use LLMs, vision models, retrieval, and evaluation workflows to extract, structure, and reason over data
- Build backend services, data pipelines, internal tools, and production systems that make extracted data available downstream
- Implement LLM-powered extraction from documents, including multilingual and historical material
- Develop vision-model-based extraction from images and documents
- Build systems that ground model outputs in trusted structured context where every claim traces to its evidence
- Implement retrieval workflows, context assembly, and answer grounding
- Create evaluation datasets, scoring workflows, and regression checks to measure quality
- Build internal tools for human review and verification of AI-generated outputs before production
- Build production-grade AI workflows with observability, latency, cost, retries, queues, and failure handling
- Develop React-based internal and reviewer-facing surfaces when needed
Requisitos
- 7+ years of professional software engineering experience
- Strong full-stack experience with a backend lean, building production services, APIs, and data pipelines
- Working React experience
- Real experience using LLMs and vision or multimodal models for data extraction, structured reasoning, classification, or entity extraction
- Experience grounding model outputs in trusted data, structured context, or retrieved information
- Demonstrated eval discipline: built evaluation datasets, scoring workflows, and regression checks
- Ability to reason about model failure modes and improve the system around the model
- Strong product and execution judgment; pragmatic version first, no over-platformizing
- Ability to work independently from problem definition to shipped system
- Production experience with TypeScript/Node, Postgres, and React
- Real experience shipping against frontier-model APIs and building evals around them
- Serverless or edge experience (Cloudflare Workers, Lambda class)
- Python for eval harnesses and data tooling, Supabase specifically
- AI-native development practice, working with coding agents
Candidate-se a vagas como esta, com um copiloto.
O JobMode.app diz se a vaga combina com você, estima a faixa salarial, adapta seu currículo e escreve respostas de triagem — direto no navegador.