AI Engineer

Applefy, a leading mobile and web development agency, is seeking an AI Engineer to join our team on an ongoing project.

The project is an AI customer-support agent for an e-commerce company. You will build the agent that answers customers from the client's own knowledge base, looks up live order and account data, and hands off to human agents in Zendesk with full context. 

This suits an engineer who wants to ship production AI, not just prototypes, and who cares about accuracy and safety as much as speed.
 

What We Offer:

• Flexible work schedule: Manage your own time and work at your own pace.
• Fully remote: Work from anywhere in the world.
• Comprehensive leave policy: All official holidays, 20 days of paid vacation, and 10 sick days without needing a doctor's note. [confirm]
• Hands-on product work: You will work directly with the founder and the client team, with weekly syncs and a demo at the end of each milestone.
   

Key Requirements:

• 4+ years of backend engineering experience, with strong Python (FastAPI or similar)
• Hands-on experience building LLM features in production, including tool calling, prompt design and output evaluation
• Practical RAG experience: chunking, hybrid search, source citations, and knowing when a model should decline to answer
• Experience integrating with third-party REST APIs, including authentication, retries and error handling
• Automated testing discipline, with unit and integration coverage targets you can explain
• A working understanding of GDPR and of protecting personal data in logs and prompts 
• Strong communication and soft skills
• Proficiency in spoken and written English
• Keen attention to detail
• A desire to learn and grow professionally
   

Preferred Qualifications:

• Experience with the Zendesk or Notion APIs
• Experience with LLM tracing and evaluation tools such as Langfuse
• Experience with prompt-injection defence or red-teaming
• Experience with EU cloud hosting and PostgreSQL with pgvector
• A React background for an embeddable chat widget
   

How to apply: Send a short note on the most relevant agent or RAG system you have built, a link to code or a demo if possible, and your availability.