AI & Machine Learning Company in Muscat

A scoped AI proof of concept at Muscat rates runs OMR 12,000 to 30,000 (roughly USD 31,000 to USD 78,000) over six to ten weeks, covering data audit, a baseline model, and a hosted demo in AWS me-south-1 Bahrain. A custom production ML model (PDO predictive maintenance, Bank Muscat fraud scoring, Madayn ICV automation, Arabic document extraction) typically lands at OMR 45,000 to 110,000 (USD 117,000 to USD 286,000) including MLOps, monitoring, and a PDPL-aligned model card. Full production AI systems with RAG, multiple models, fine-tuning, and SAP or Maximo integrations range from OMR 95,000 to 420,000 (USD 247,000 to USD 1.1M). Muscat rates sit below Dubai and Abu Dhabi but reflect the bilingual and ICV documentation overhead. We provide fixed-fee proposals rather than open T and M estimates.

Royal Decree 6/2022, which became enforceable in February 2023, is Oman's first comprehensive personal data law. It requires a lawful basis for processing, explicit consent for sensitive categories (health, biometrics, genetic, religious belief), a Ministry of Transport, Communications and Information Technology permit for cross-border transfers outside the GCC under most conditions, and breach notification. Our discovery phase runs a PDPL classification on every data set and use case, and high-impact builds ship with a Record of Processing Activities, a Data Protection Impact Assessment, and a residency plan that keeps personal data inside AWS me-south-1 Bahrain or an in-country host. Penalties under the law reach OMR 500,000, so Bank Muscat and PDO already require PDPL-grade governance regardless of project size.

Yes. We ship reservoir analytics, predictive maintenance on rotating equipment, anomaly detection on SCADA and historian streams, flare and methane computer vision, and ESG and emissions forecasting tied to Vision 2040 net-zero-by-2050 commitments. Our deployments respect Purdue model segmentation, sit on the IT side of the DMZ with read-only access into PI Historian or AVEVA, and feed existing CMMS systems (Maximo, SAP PM) rather than displacing them. For PDO's ICV obligations we structure subcontracting, training, and local hire so the scope contributes to your ICV score, and we deliver Arabic operator-facing dashboards alongside English engineering documentation.

Yes. We ship Mudarabah profit-sharing scoring, Murabaha cost-plus financing risk models, Ijarah leasing portfolio analytics, AML transaction monitoring tuned for Omani remittance corridors (India, Pakistan, Bangladesh, Philippines, Egypt), KYC document AI for Omani civil IDs, and Arabic-Latin name matching for sanctions screening. Models are built under the Central Bank of Oman cybersecurity framework, with model risk documentation that mirrors Basel SR 11-7 patterns and respects Shariah board review cycles. We have patterned governance after public Bank Muscat and Bank Dhofar disclosures, including challenger model testing, drift monitoring, and human-in-the-loop overrides on every adverse action.

Yes, and we treat it as a first-class concern, not a translation afterthought. We work with Jais (the Inception G42 Arabic LLM), Cohere multilingual, Anthropic Claude, and GPT-4-class models with Arabic prompts, plus open-weight Llama 3 and Mistral fine-tuned on Modern Standard Arabic and Gulf dialect data. For OCR on Omani civil IDs, driving licences, and Arabic invoices we combine Tesseract Arabic, AWS Textract Arabic, and custom CRNN models. RAG retrievers use mE5, BGE-M3, and Cohere multilingual embeddings tuned on Omani Arabic corpora. Evaluation harnesses run blind Arabic and English test sets before every production push, so accuracy is measured in both languages and reported separately.

We default to AWS me-south-1 Bahrain for training, storage, and inference because it is the closest stable AWS region with full Bedrock, SageMaker, and Textract Arabic availability. Azure UAE North (Dubai), Oracle Muscat region, and OQ Telematics-adjacent Omantel data centres serve as alternates when client policy demands GCC-resident or in-country compute. For LLMs that must stay regional we self-host Llama 3, Mistral, or Jais on Gulf GPU instances. PDPL cross-border transfers outside the GCC trigger a Ministry permit, so we document residency in the model card and the data processing agreement up front and avoid US or EU regions unless your DPIA explicitly clears them.

A typical Muscat production model takes sixteen to twenty-four weeks from kickoff, assuming clean historical data and PDPL and CBO documentation are inside scope. Week 1 to 4 is data audit, PDPL classification, and baseline modelling. Week 5 to 12 is iteration, Shariah review where relevant, and challenger testing. Week 13 to 18 is MLOps, drift monitoring, shadow-mode deployment, and internal audit walkthrough. Week 19 onward is staged rollout. PDO and OQ digital twin projects often add four to eight weeks for OT integration and Purdue-model review. If you are pre-data or need Arabic labelling, add six to ten weeks. We have shipped to this cadence inside CBO-supervised and ICV-graded environments.

ICV is Oman's local content framework, originally PDO-led and now applied across the Ministry of Energy and Minerals and most state-linked procurement. Vendors are scored on local goods, local services, Omanisation of staff, training spend, and supplier development. Our Muscat engagements deliver an ICV impact statement up front, structure subcontracting through Omani-registered SMEs where it materially improves the score, and route training hours through SQU, GUtech, or Madayn-affiliated programmes when scope allows. We are transparent that as a regional delivery partner we do not maximise ICV the way a fully Omani-owned firm does, but we structure engagements so the AI line item contributes positively rather than dragging your composite score.