Meana Petro
KC-02FrontierWave 3AI

Domain-Specific Language Model

In one sentence

A model further trained on petroleum engineering text, so it uses the field’s vocabulary correctly in Persian and Arabic as well as English.

The problem

General-purpose models handle petroleum terminology unevenly and are notably weaker in Persian and Arabic technical registers, where they translate terms of art literally and produce text a domain specialist immediately distrusts.

How it works

Continued pre-training and supervised fine-tuning on curated petroleum engineering corpora across the four languages, with the terminology table used as a consistency constraint. Evaluation is against domain benchmarks authored by practising engineers, not against generic language benchmarks.

Data required

  • Curated domain corpus with clear rights to use it for training
  • Terminology table across the four target languages
  • Expert-authored evaluation sets per language
  • Substantial GPU capacity for training and evaluation

What you get

  • Domain-adapted model deployable on premise
  • Benchmark results per language against the base model
  • Terminology consistency report across the four languages

What it cannot do

This is frontier work with a large cost and an uncertain margin over a well-prompted general model plus good retrieval. We recommend it only where terminology consistency in Persian or Arabic is a hard requirement, and we will say when retrieval alone is the better investment — which today it usually is.

Every capability page states a real limitation. If a vendor cannot tell you where their product stops, you are not being told the truth about where it starts.

Related capabilities

  • KC-01Foundational

    Technical Knowledge Assistant (RAG)

    Ask a question about your own field in Persian, Arabic, English or Spanish and get an answer with the report and page it came from.

    Knowledge & CopilotAI
  • DF-02Foundational

    Structured Extraction from Reports

    Pulls the numbers out of the reports — depths, rates, pressures, volumes, dates — and puts them into a database with a link back to the page they came from.

    Data FoundationAIDomain
  • KC-03Frontier

    Multimodal Subsurface Assistant

    Understands log images, maps and cross-sections alongside text, so you can ask a question about a figure rather than only about a paragraph.

    Knowledge & CopilotAI