PromptHeroes.ai logo

AI Engineer - Agents & Harnesses

PromptHeroes.ai

Oostzaan
Full-time
2-5 years experience
On-site

Key Skills

Python
AI Agent Development
Agent Orchestration
Tool Calling
MCP Integrations
API Integration
Prompt Injection Security
Least-Privilege Access
Sandboxing
Secrets Management
LLM Evaluation
Regression Testing
RAG
Embeddings
Git
Docker

Job Description

About PromptHeroes PromptHeroes builds AI solutions for companies: agents and automated workflows that take real work off people's hands, connected to the data and systems they already use. We also build our own AI products and start-ups, and we create content with generative AI for brands and agencies. Our core is a team of English-native developers and prompters in Nairobi, Kenya, working with our team in Amsterdam and Oostzaan for clients in the Netherlands. The role You build the systems that let AI agents do real work, reliably and safely. That means the harness around the model: orchestration, tool calling, MCP integrations, state and retries, schema-validated outputs, and the connections to client data and APIs. You build them secure by design: agents only get the access they need, are hardened against prompt injection and data leakage, and every change is tested against evals before it goes live. You work on client projects and on our own ventures, with our teams in Oostzaan and Nairobi, and you talk directly with clients. Full-time, on-site in Oostzaan. What you bring Strong Python, plus experience building on the APIs of OpenAI, Anthropic, Google or open-source models You have built and shipped agent systems: planning, routing, tool calling, structured outputs, fallbacks when a provider goes down You build with security in mind: prompt injection, least-privilege permissions, sandboxing, secrets management, and testing how your own agents can be misused Evaluation as a habit: reference sets, LLM-as-judge, regression tests, comparing models on quality and cost before switching RAG and embeddings, and enough ML to read a fine-tune and judge whether it worked Comfortable with Git, Docker and a cloud platform (AWS, GCP or Azure) You can explain to a client in plain language what an agent does, and what it is not allowed to do What you get Market-conform salary with the possibility of equity. Highly Skilled Migrant visa sponsorship is available for the right candidate. Challenging projects for clients and for our own ventures, in a small team that moves fast. Send [email protected] a short motivation and two or three things you built.

Core Responsibilities

Build reliable, secure AI agent systems, including orchestration, tool calling, integrations, state management, retries, and schema-validated outputs. Work on client projects and company ventures, test changes with evaluations, and explain agent capabilities and limits directly to clients.

Requirements

Candidates should have strong Python skills and experience building and shipping agent systems using AI model APIs, with knowledge of security practices such as prompt-injection defenses, least-privilege access, sandboxing, and secrets management. They should be comfortable with evaluation and regression testing, RAG and embeddings, Git, Docker, and a cloud platform, and be able to communicate clearly with clients.

Benefits

  • Equity

About PromptHeroes.ai

Industry: Technology, Information and Media

Company size: 11-50 employees

PromptHeroes is an AI first agency that utilizes relevant AI tools to streamline the content production process for high end output at scale. Also our developers help automate and develope your tech projects. Our unique advantage lies in our community of tech-savvy, well-trained, English-native prompters and developers in Nairobi, Kenya. These skilled talents deliver quality content, technology and agentic services for brands and agencies in the Netherlands.

Added Today