Senior AI Java Engineer

Forte Group; Colombia; On-site; COP 120,000,000 - 240,000,000; Full time

Job summary

Forte Group is helping a US fintech client make AI part of how they build software, from idea to prototype to production-ready code. You will be the technical lead: design and build the AI layer that connects the client's knowledge to their delivery process, reusable skills and agents for each step, and the platform to share and govern them across teams.

The role emphasizes hands-on design and shipping, with collaboration from senior architects, whiteboarding in the morning and shipping by the

Qualifications

• 5+ years in software engineering; strong Java and Spring Boot; working TypeScript (React or Node.js).
• 1+ year building with LLM APIs: tool use, structured output, prompt design, agents.
• Agent-based coding beyond autocomplete (Claude Code, Cursor, Copilot agent mode): custom skills, rules, MCP.
• Git, CI pipelines and automated testing (JUnit or similar).
• REST APIs and system integrations; Python scripting.
• Self-starter who ships small, gets feedback and iterates.
• Works directly with client engineers and explains clearly; fluent English (B2+/C1).
• Quality-minded: tests what they build and checks AI output instead of trusting it.

Responsibilities

• Build and maintain Claude Code skills, subagents, hooks, commands and plugins for product, design, architecture and engineering workflows.
• Build MCP servers and connectors to Confluence, Jira, Git and Figma, with retrieval that returns the right content and cites its source.
• Build code and unit-test generation that follows the client's Java/Spring Boot service patterns and coding standards, with lint and tests wired into the loop.
• Build flows that turn product ideas into clickable React/TypeScript prototypes in the client's design style.
• Write evaluation sets and regression checks for every skill; track quality, latency and token cost.
• Pair with pilot-team engineers, collect feedback, fix what gets in their way, and document how to use what you build.

Skills

Java

Spring Boot

TypeScript

React

Node.js

LLM APIs

Agent-based development

REST APIs

Python scripting

Git

CI pipelines

Automated testing

Tools

Claude Code

Cursor

Copilot

MCP servers

Confluence

Jira

Figma

JUnit

Job description

Forte Group is helping a US fintech client make AI part of how they build software, from idea to prototype to production-ready code. You will be the technical lead: design and build the AI layer that connects the client's knowledge to their delivery process, reusable skills and agents for each step, and the platform to share and govern them across teams. The client's architects are strong and already build with AI, so you earn trust by building, not by presenting. Expect to whiteboard a design with senior architects in the morning and ship it in the afternoon.

What you'll do

• Build and maintain Claude Code skills, subagents, hooks, commands and plugins for product, design, architecture and engineering workflows.
• Build MCP servers and connectors to Confluence, Jira, Git and Figma, with retrieval that returns the right content and cites its source.
• Build code and unit-test generation that follows the client's Java/Spring Boot service patterns and coding standards, with lint and tests wired into the loop.
• Build flows that turn product ideas into clickable React/TypeScript prototypes in the client's design style.
• Write evaluation sets and regression checks for every skill; track quality, latency and token cost.
• Pair with pilot-team engineers, collect feedback, fix what gets in their way, and document how to use what you build.

Must have

• 5+ years in software engineering; strong Java and Spring Boot; working TypeScript (React or Node.js).
• 1+ year building with LLM APIs: tool use, structured output, prompt design, agents.
• Agentic coding beyond autocomplete (Claude Code, Cursor, Copilot agent mode): custom skills, rules, MCP.
• Git, CI pipelines and automated testing (JUnit or similar).
• REST APIs and system integrations; Python scripting.
• Self-starter who ships small, gets feedback and iterates.
• Works directly with client engineers and explains clearly; fluent English (B2+/C1).
• Quality-minded: tests what they build and checks AI output instead of trusting it.

Nice to have

• Has built MCP servers; Claude Agent SDK, LangGraph or similar.
• Retrieval and vector search; knowledge graphs.
• Event-driven architecture or DDD.
• Contract testing and mocking (Pact, WireMock).