Get in touch
Data science & AI · Sharjah, UAE

Turn the data your business already has into decisions and automation.

Kalson builds custom data science and AI systems on top of your spreadsheets, sales records and documents, and sets them up where your team already works.

Get in touch See real projects
  • Fixed price, agreed upfront
  • Built by the engineer you talk to
  • You own the code
  • In person in the UAE, remote worldwide
What you have
  • SpreadsheetsExcel, Sheets
  • Sales and ordersPOS, Shopify, Salla
  • Accounting, ERPTally, Zoho, Odoo
  • PDFs and scansinvoices, forms
  • Photosproducts, sites
  • Laptop-only toola model or a script
  • Ideas, no plan yetwhere could AI help?
Examples of what you get
What Kalson does

Which of these sounds familiar?

Seven problems growing companies run into, and what Kalson builds to fix each one.

Data analysis and dashboards

“Someone rebuilds the same report by hand every week, and nobody agrees on which number is right.”

One trusted view of sales, stock, finance and operations, refreshed automatically.

Machine learning and forecasting

“Stock, staffing and purchasing are planned on gut feeling.”

Forecast demand, spot customers who are about to leave, catch unusual transactions, and set prices with evidence instead of guesswork.

AI and LLM applications

“Invoices and forms are typed in by hand, and our team answers the same questions all day.”

Assistants, document processing and internal search built on your own documents.

Computer vision

“People check photos, scans or products by eye, which is slow and inconsistent.”

Models that inspect, classify and sort images, and highlight what each decision was based on.

Data pipelines and automation

“Our data sits in five systems that don’t talk to each other.”

Connect your ERP, CRM, spreadsheets and APIs so data flows without copy and paste.

Deployment and maintenance

“There is a model or a script that works on one laptop, and nobody else can use it.”

Turn a working model or script into a secure tool your whole team can use, and keep it running.

Data and AI consulting

“We know data and AI could help, but not where to start or what is worth paying for.”

A clear, costed plan before you build anything: what to do first, what it is worth, and what to skip.

Not sure which one you need?

Many problems need two or three of these together. Describe the problem in your own words and I will suggest where to start.

Describe your problem
Every service in detail
Projects

Real systems, built end to end.

Three complete systems, from raw data to a working app. Every demo is recorded from the running app, and every number comes from the project itself.

Computer vision · Explainable AI

RadiantCareResearch prototype

For research use only

Chest X-ray AI that highlights the image regions behind its prediction and writes a plain-language report

14 findings
scored on each chest X-ray
86,000
X-rays used to train the model
≈0.2 s
for the model to score one X-ray on an ordinary processor, no graphics card needed
The same approach in a business
  • Spotting defects on a production line
  • Checking site, shelf or vehicle photos
  • Reading and sorting scanned documents
See the full project
Data pipeline · Dashboard

PostOilIndependent project

A live dashboard of how the six GCC economies are moving beyond oil

6 economies
Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, UAE
10 sources
joined into one dataset, each figure with its year and source
Live
rebuilt and republished automatically on every update
The same approach in a business
  • Management dashboards across branches and systems
  • Automated weekly and monthly reporting
  • Market and competitor data for GCC expansion
See the full project
AI for documents · LLM application

EducAIJoint Master’s project

Built with Kamul Ali Nassoma Wattara

An AI assistant that writes maths exercises and marks pupils’ typed PDF answer sheets

Auto-marking
pupils’ typed PDF answer sheets read and marked against the answer key
5 areas
of the maths curriculum: numbers, calculation, measurement, geometry and logic
Teacher signs off
any mark can be changed before results are saved
The same approach in a business
  • Invoices and purchase orders checked against contracts
  • Insurance claims and application forms
  • Compliance checklists and audits
See the full project
Industries

Ideas for a first project in your sector.

For companies with 20 to 500 people that have data but no data team.

Logistics and freightRead shipping documents automaticallyRetail and e-commerceDemand and stock forecastingReal estate and property managementLead scoringClinics and healthcare groupsPredict no-showsManufacturing and tradingVisual quality checksHospitality and restaurantsReview analysisProfessional servicesSearch and summarise your documentsSchools and training centresExercises and first-pass marking
Another sector? Tell me about your data
How it works

Start small, prove it, then scale.

Four steps from the first message to a working result, with a fixed price agreed before each paid stage starts.

  1. A free first conversation

    Write to me through the form, by email or on WhatsApp, and we talk it through in writing or on a call. You get a straight answer on whether data science or AI will help.

    Get in touch

  2. A fixed-price proposal

    A short written proposal: the scope, what is not included, the milestones, the timeline and a fixed price for each stage.

  3. A pilot on your real data

    We start with a small, paid first step that proves the idea on your own data, usually in a few weeks. For example, a sales forecast for one product line before the whole range.

  4. Build, deploy and hand over

    If the pilot works, the full system is built: the same idea extended to everything in scope, deployed where your team uses it and documented. You own the code once it is paid for.

More on how projects work
Why Kalson

Direct access, honest scoping, no surprises.

“I started Kalson so that growing companies can get serious data science and AI work without a consultancy budget, from the engineer who builds it and answers for the result.”
Sulaiman Fayyaz · Founder, Data Scientist & AI Engineer
More about Sulaiman

  • You talk to the engineer

    No account managers. The person on your call is the person who designs, builds and deploys your system.

  • A licensed UAE company

    You sign a written contract with Kalson FZE, licensed in the Sharjah Publishing City Free Zone, and pay against proper invoices: part upfront, the rest at milestones you can see.

  • You pay for what you need

    Some projects need only an analysis or a trained model, others a full system deployed and maintained. The proposal lists exactly what is delivered, and each stage has a fixed price agreed before it starts.

  • You own what you pay for

    Once a project is paid in full, the code and the work belong to you. No lock-in, no licence fees to me.

  • Confidential by default

    Your data is used only for your project, then returned or deleted. I will sign an NDA before you share anything.

  • Honest about AI

    I measure results on data the model has never seen and tell you the limits. If AI is not the answer, you will hear it from me first.

Questions

Good questions to ask before you start.

All questions
How much does a project cost?

A fixed price for each stage, agreed in writing before that stage starts. It depends on what you need: an analysis, a trained model or advice costs far less than a complete system that is deployed and maintained, and you pay only for what the proposal lists. The first step is a free conversation.

How can we judge your work before starting?

Look at the projects: each one has a demo recorded from the running app, the code and the results. Then start with a small fixed-price pilot on your own data, so you see real results before committing to more. Payments follow milestones you can check, and every project has a written contract with Kalson FZE, a licensed UAE company.

What if the pilot does not give good results?

Before the pilot starts, the proposal states what a good result means for you, such as forecast accuracy, hours saved or the share of documents read correctly. If the pilot falls short, you have paid for a small pilot, not a full system, and you get a written summary of what was tested, what your data showed and whether another approach is worth trying. You are never committed to the next step.

Do we need a lot of data?

Less than you might think. Many useful projects start with a year or two of sales, a folder of PDFs or an export from your ERP. On the first call I will look at what you have and tell you whether it is enough.

Can you work with our existing systems?

Yes. I can work from exports or APIs of ERPs and CRMs (Odoo, Zoho, SAP Business One, Microsoft Dynamics and others), online stores, point-of-sale systems, databases, spreadsheets and PDFs. The goal is to fit into how your team already works.

Can the AI work in Arabic?

Yes. Current language models read and write Arabic well, so documents, assistants and product content can be bilingual. I test on a sample of your real Arabic material before promising any accuracy.

Have data that should be doing more work?

Tell me what slows your team down, on a call or in writing. You will get an honest answer on whether data science or AI can fix it, and what a first fixed-price step would cost.

Send a message

Free first conversation, no commitment.