Altum Software is looking for a Prompt Engineer for whom "close enough" is never close enough.

This isn't an "AI enthusiast" role. It's a hands-on position where you design, test, and iterate on prompts that directly shape the quality of AI-powered products — for our clients and our own internal workflows.

If you like finding the exact phrasing that makes a system behave correctly — and you're obsessive about the difference between "almost right" and "exactly right" — read on.

Who we are: Altum is a Ukrainian software and website development house. We build human-centered digital products for startups and mid-sized businesses, and support digital agencies from the US and Western Europe. No mediocrity, no passengers — only people who take ownership and care about doing things well.

What we offer:

• Direct exposure to real AI product work — not demos, not prototypes in a vacuum;
• A team that actually ships things, fast;
• Mentorship and continuous knowledge exchange;
• Access to modern AI tools and platforms;
• Clear room to grow — as the field grows, so does the role;
• Regular, honest feedback;
• Paid time off and sick leaves;
• Personalized benefits built around you.
  

What you need:

• A genuine obsession with language, logic, and how models interpret both;
• Strong written English — you'll be prompting, documenting, and communicating in it daily;
• Understanding of how LLMs work — not theoretically, but practically: context windows, system prompts, temperature, hallucinations, edge cases;
• Experience writing and testing prompts across different models (GPT, Claude, Gemini — any);
• Structured, analytical thinking — you can break down a vague problem and build a precise solution;
• Ability to evaluate outputs critically, not just by "feeling";
• Familiarity with tools like LangChain, PromptLayer, or similar — a plus;
• Background in linguistics, technical writing, UX writing, or software development — also a plus.
  

What you'll do:

• Design and refine prompts for AI features across client products and internal tools;
• Build and maintain prompt libraries — structured, versioned, reusable;
• Test systematically: document inputs, outputs, failure cases, and improvements;
• Work closely with developers, PMs, and designers to integrate AI into product flows;
• Research and apply new techniques — chain-of-thought, few-shot, RAG-adjacent approaches;
• Translate vague business requirements into precise model instructions;
• Track how prompts perform over time and iterate based on results.