AI/LLM Engineer

Location: Remote, global; the department is listed as Poland (Remote).

Company

hirify.global develops connectivity and distribution products with compute capabilities and an AI concierge called YOlanda.

What you will do

• Own YOlanda’s system prompt, persona, tone, safety rules, refusal behavior, handoffs, and prompts for V3 actions including plan changes, top-ups, balance queries, offers, troubleshooting, and account support.
• Design conversation flows that reach a real product or service within five responses, manage context, break tasks into chains and tool calls, and define the boundary between YOlanda and YOnC.
• Build bilingual evaluation sets with test cases, pass conditions, scoring, boundary measurements, fact checks, and red-team tests for top-up and rewards flows.
• Trace calls and operate production systems covering prompt versions, models, tokens, cache hits, latency, cost, SLOs, alerts, timeouts, retries, and provider or model fallbacks.
• Build semantic, retrieval, and tool-result caching; optimize token budgets, context trimming, output length, batch APIs, model routing, accuracy, latency, and cost per turn.
• Implement version control, CI release gates, canary and staged rollouts, kill switches, audit logging, personal-data redaction, prompt-injection defenses, tool permissions, spend limits, human approval points, and cost budgets.

Requirements

• At least two years of experience building and operating large language model systems in production at real traffic.
• Shippable Python skills, including scripts, API clients, data handling, and tests; Git, JSON, regex, and Markdown.
• Production experience with LangChain and LangGraph, including changing an existing graph without breaking it.
• Shipped system prompts and evaluation harnesses with controlled, one-variable-at-a-time testing and recorded results.
• Production experience with tracing, SLOs, fallbacks, and correctly invalidated caching, including measured cost and latency improvements.
• Practical governance experience with versioning, release gates, audit logging, personal-data handling, and controls implemented in code.
• Production retrieval experience with chunking, embeddings, reranking, grounding answers in sources, and identifying when retrieval is inappropriate.
• Exceptional written English with strong syntax, register, and nuance; ability to write in a voice other than your own.
• Care for users on prepaid plans and awareness that incorrect answers can cost users money.

Culture & Benefits

• Highly collaborative team with a shared drive to achieve goals.
• Flexible working hours.
• Competitive salary.
• Opportunities for personal and professional growth.
• Work on products disrupting the traditional media-services industry.