We are looking for an AI engineer with 4-8 years of experience to work on our agentic product.

Salary 75-90k EUR (depending on seniority)

Equity 0.5-1%

We are Perry and we build work instructions, and validate work quality for technical operators using AI. We are VC funded by Revent (Berlin) and angels from OpenAI, Google and other top companies.

The company is founded by Laurens (ex-Google) and Arie (ex-McKinsey) because we solve a big societal issue: most developed economies will come to a standstill if we don't improve the efficiency of technical operators, especially in the energy transition. We use AI to rapidly build technical work instructions that are delivered to the operator in a way that actually works in the field.

Take a look at what we do on perry.works

One of the key components of our product is Perry Brain: the data pipeline and multi-agent AI that creates work instructions and validates work quality. Our vision is to build a live video conversational agent that deeply understands technical work and helps instruct and troubleshoot. Getting this right is the key thing that allows the company to succeed. We are expanding the team with a top-tier AI engineer. 

You will work primarily on our agentic system (Perry Brain) and the back-end of Perry Studio (where we create, manage & track the instructions and work). However, you are expected to take a full-stack/full-product perspective, including helping take action on the whole product, 

What you can expect:

• Work at the cutting edge of AI/LLM model evaluation, agent orchestration and context engineering; we continuously experiment with newer models/frameworks and build our product with upcoming models in mind. "If your product is replaced by an LLM in 12 months, you're building the wrong product"
• A rapid learning environment: You work very closely with the founders, our other AI engineer, front-end engineering and product designer. We're moving fast, and we go deep.
• Be part of our success early on and build and grow the company together with us 

What we expect from you:

• You’ve got LLM-in-production experience: you have experience with productionize LLMs in production environments and understand the importance of context and evaluation of LLM outputs.
• You have built RAG pipelines and evaluation at scale
• You’re already an advanced user of coding agents and understand their limitations: you have extensive experience developing using coding agents like Claude Code, Cursor, Codex or others and come up with new ways to accelerate your own (and our) engineering speed, while knowing today’s limits (no vibe coders)
• You have high ambition to build Perry into a rapid growth company with us. Let us be clear; this is not a 9-5 job, but it will be mega rewarding.
• You take radical ownership, not just for the part you are working on but also for the entire company and team.

Nice to have:

• Already familiar with some of our key technologies such as Langgraph, Pydantic(AI) and various LLM/VLMs/OCR models. You should be able to reason through technology selections for the right use cases. 
• Advanced agentic RAG techniques such as agentic RAG, contextual RAG, hybrid RAG, etc
• LLM model training, finetuning, open source model deployment
• You understand various AI vision techniques and have trained vision models

Our culture

Culture is very important to us and it is important that you not only subscribe to it, but also feel you can live this for the next employees. Our values:

• Strong opinions, loosely held
• Disagree and commit
• Integrity over money and success
• Be scrappy, find a way
• Make the other person shine
• Create impact

Some of the technologies we use:

On the back-end, we are building a Multi-Agent AI framework with orchestrator agents and subagents. Some key technologies we use:

• Python
• React
• LangGraph
• Pydantic(AI)
• Supabase
• FastAPI
• Various models : GPT-5; Claude Sonnet/Opus; Gemini Pro/Flash; Mistral
• Various agentic RAG approaches (Pinecone) and knowledge graphs (Neo4j)