AI Engineer

Who are we?

Sempico Solutions is an international telecom company and business messaging service built around our own product — the Gatum messaging platform. We help our clients deliver different types of messages, such as marketing campaigns, verification messages, OTPs (one-time passwords), and more.

Our mission is to provide fast, simple, and reliable communication between businesses and customers — through SMS and messaging apps — with global reach, anywhere, always.

Our benefits:

• Work in a fast-growing product company where you can grow together with the business.
• Annual paid vacation of 24 calendar days and paid sick leave.
• Remote working format.
• Convenient working hours from Monday to Friday from 9:00 to 18:00 UTC+2 (including 1 hour for lunch).
• Competitive salary: base rate with an opportunity to review each 6 months.
• Step-by-step onboarding and training process.
• Planning your individual career path together with your manager.
• Reimbursement of expenses for professional development courses/certifications (up to 75% in agreement with the management).
• Strong positive product-company culture with a good sense of humor.

What about the team?

🏡 100% remote | 🌎 Global team

At our company, everyone is on a first-name basis. There's no unnecessary bureaucracy. We work together, support each other, and move toward common goals. At Sempico Solutions, you'll quickly feel what it means to be a real part of the Team.

Requirements

At Sempico Solutions, we're looking for a hands-on AI Engineer (Middle+/Senior) who doesn't just experiment with models — but builds AI systems that work reliably in production and deliver measurable impact on real business processes.

You are someone who thinks in terms of quality metrics, not "generally good." You bring structure to data and logic, take ownership of AI outcomes, and know how to move from idea to stable production without losing quality along the way.

Our ideal candidate has:

• 3+ years of experience in AI/ML in production environments (not pet projects).
• Confident Python skills — your primary working language.
• Hands-on experience with LLM APIs in production (OpenAI, Anthropic, or open-source models).
• Systematic approach to prompt engineering: versioning, A/B testing, quality control.
• End-to-end RAG architecture experience: chunking, embeddings, retrieval, reranking, vector databases (Pinecone, Qdrant, Weaviate, or pgvector).
• Experience building AI evaluation frameworks — automated quality tests for AI responses.
• Knowledge of privacy and GDPR considerations when working with client data.
• REST API integration experience and async I/O work.
• Basic DevOps skills: Docker, monitoring, logging.
• Fluent Ukrainian (C1 or native) and English at B2 Upper Intermediate level or higher (both spoken and written).
• Understanding of prompt injection risks and data leak prevention.

Nice to have:

• Experience integrating messengers: Telegram Bot API, WhatsApp Business API, Microsoft Teams Graph API.
• Experience with support platforms: HelpCrunch, Intercom, or Zendesk API.
• Familiarity with LangChain, LlamaIndex, or Haystack.
• Classical ML background: scikit-learn, PyTorch, or TensorFlow.
• Experience in CPaaS / SMS industry.
• Automation tools: n8n, Zapier, or Make — for rapid flow prototyping.
• PHP reading ability (to navigate the existing product codebase).
• Experience building AI systems from scratch in a startup environment.

What You'll Do

• Build and maintain a unified client inbox — bringing all communication channels (Telegram, WhatsApp, Teams, HelpCrunch, email) into one structured place.
• Launch and continuously develop the AI Support mechanism across all client chat channels.
• Form and maintain the knowledge base that powers AI responses.
• Monitor AI response quality in production: track metrics, run A/B tests, proactively detect degradation.
• Control and optimize AI call costs — you know what your solution costs per month in tokens.
• Integrate AI into other business processes: sales, analytics, and internal tooling.
• Document AI architecture, flow logic, and known limitations clearly — for both technical and non-technical colleagues.