AI Engineer

ABOUT WELVAART

On a daily basis, we assume commitments and present solutions to our stakeholders in order to create a structure of human values, based on professionalism, honesty and rigor.

With a management based on Human Centered Design, we take care of our professionals with consistent career plans, but flexible with their needs and expectations of evolution. Our management team guarantees an empathetic and present leadership that will provide superior technological engagement and delivery to our clients' projects and products.

Project

A project within a newly created squad of five engineers responsible for maintaining and evolving production AI agents used by the back office of the securities services (custody) division of a major European bank. The agents are built in Python using CrewAI, with a planned migration to an internal SDK and LangGraph. Surrounding APIs are developed in Go, and all agents are containerized with Docker and deployed on Kubernetes. The squad operates within a federation of local AI teams supported by a central AI group delivering general‑purpose agents that must be adapted to specific custody workflows.

Role

• Design, build and continuously improve AI agents, including reasoning flows, tool usage, prompts and evaluation


• Adapt central general‑purpose agents (drafting, summarization, process orchestration) to custody back‑office use cases


• Develop and maintain agents in Python using LangGraph, LangChain and the internal SDK


• Design agent workflows: task decomposition, tool/function calling, memory and state handling, guardrails and human‑in‑the‑loop checkpoints


• Engineer, test and version prompts; manage configuration changes through Kubernetes ConfigMaps


• Build and run agent evaluation using golden datasets, regression suites and quality metrics


• Localize central agents by analysing gaps between general‑purpose behaviour and local business requirements


• Instrument agents for observability: reasoning traces, token usage, latency and failure modes


• Manage model lifecycle concerns: provider changes, context window constraints, cost/latency trade‑offs


• Work daily with the Business Analyst to translate operational knowledge (settlements, corporate actions, client queries, reconciliations) into agent behaviour

• 3–6 years of Python engineering experience, including 1–2 years in LLM/GenAI applications


• Practical experience with agent frameworks (CrewAI, LangGraph, LangChain, AutoGen or similar)


• Understanding of ReAct‑style loops, tool calling and structured outputs


• Strong knowledge of prompt engineering, RAG patterns and LLM evaluation techniques


• Experience consuming REST/gRPC APIs (Go backend), JSON schema design and async Python


• Working knowledge of Docker and Kubernetes (deployments, logs, ConfigMaps)


• Strong testing culture: pytest, mocking LLM calls, deterministic test design for non‑deterministic systems


• Strong analytical and problem‑solving skills


• Good communication and teamwork abilities

• Be part of a tech start\-up


• Different scopes of project in different sectors


• Structure of fairness and equity salary (Consultant Profile)


• Training \& Certification


• Career Path management


• More than 30 Partnerships


• Welvaart Ambassador Program

Pagamento: Até 2 700,00€ por mês

Localização do trabalho: Trabalho remoto híbrido em Lisboa

Este anúncio é de Indeed. Ver anúncio original ↗