Frequently asked questions

Yes — it is the centre of what we do. We build LLM training and fine-tuning, local model deployments, RAG systems, AI agents, voice AI call centers and process automation. From assistants grounded in your knowledge base to end-to-end workflow automation integrated with your existing systems.

RAG adds knowledge, fine-tuning teaches behaviour. If the model needs current facts from your documents, that is RAG; if it needs to apply your tone, output format and decision logic consistently, that is fine-tuning. We usually recommend starting with RAG — it is faster, cheaper, and updates automatically when a document changes.

It does not have to. We can run models on your own servers or inside your own cloud account, in which case data never leaves the organisation. If a commercial API is used, we document exactly which data goes where and design the setup around KVKK and GDPR requirements.

For a narrow, well-defined task with LoRA, a few hundred to a few thousand high-quality examples is often enough. Broad domain adaptation or continued pretraining needs raw text in the millions of tokens. We look at what you have and give a realistic number in the first call.

Yes. We integrate capabilities such as smart search, summarisation, document processing, conversational support and automatic classification through secure APIs, without disrupting your current stack. A rewrite is rarely necessary.

The training outputs, adapter weights, datasets and scripts are yours. The only constraint is the base model's own licence; we review the terms of families such as Llama and Qwen together at the start of the project.

We provide updates through regularly scheduled meetings. In them we go over milestones, timelines and any adjustments, so you always know where the project stands.

Typically we schedule weekly meetings to review progress, answer your questions and keep the work aligned with your goals.

We work hard to meet the dates we give; if something slips, you will not hear it late. We inform you immediately and bring a plan to get the project back on track.

For international projects we take payment through Upwork, which gives both sides a secure and transparent process. For corporate projects in Türkiye, direct contracting and invoicing is also possible.

Payments are usually structured as milestones, with each becoming due as that stage is completed and approved.

All costs are set out explicitly in the contract. For AI projects we also share an up-front estimate of monthly model/API and infrastructure costs, so there are no surprise line items.

We offer ongoing support packages tailored to your needs: technical support, updates and maintenance. Fixing issues not caused by us, and adding new features, are charged separately.

We aim to respond to every support request within 24 hours, with expedited support available for urgent issues.

It is your choice. We can hand the system over to your team with training, or keep running it under an operating agreement covering monitoring, model updates, cost tracking and quality measurement.