Book a consultation

Data and AI engineers in Tashkent, building for production

DATAI Services designs, builds and runs data platforms and AI systems for companies in Europe, North America, the Gulf and Central Asia.

Based in
Tashkent
Working hours
US time zone
First call
30 minutes with a senior engineer
Ways to work
Project team or embedded engineers

Why DATAI exists

We build data platforms and AI assistants that still have to run a year after launch.

DATAI is a team of data and AI engineers working from Tashkent. We build pipelines, lakehouses, analytics and AI systems for international companies, and we build them to run in production: scheduled, tested, monitored and documented, in the client’s own cloud.

The hard part comes months after the demo. A source renames a column and a dashboard quietly goes wrong, the person who wrote the jobs moves on, or finance asks what each AI answer costs. We plan for those problems from the start, because none of them is cheap to fix after launch.

So the way we work is written down. Code goes into your repositories. A senior architect reviews every merge request. Every pipeline has tests and monitoring from day one, and you see working code in a demo every week. At the end, your team gets runbooks and recorded training, or we keep running the platform with you.

That standard is also how we train people. New engineers join through our internship and learn it on real work, with the same review as everyone else.

Mission and values

Our mission is to give companies data and AI systems they own outright and can trust in production. Each value below is a practice you can check, and each one says where.

  1. You own what we build

    Code lives in your repositories and data stays in your cloud, in open formats. Your team can run what we deliver without us.

    Where to check

  2. A senior architect reviews every change

    A senior architect reviews every merge request, including work from engineers placed inside your team and from our interns.

    Where to check

  3. Broken data raises an alert

    Every pipeline runs schema, freshness and volume tests, and alerts go to the owning team. AI releases have to pass an evaluation set before they ship.

  4. You see progress every week

    We send a written plan before the build starts, then demo working code in your repositories every week, so you can judge the work rather than a status report.

    Where to check

  5. Access you can see and revoke

    We sign an NDA before any access. Our engineers then work through least-privilege accounts in your identity system, one per person.

Technical expertise

Six service lines across seven platform layers, from cloud infrastructure to AI in production. Each line runs on the same delivery standard.

Listing a technology does not imply a vendor partnership or certification.

The stack, layer by layer

  1. Data Engineering

    Batch and streaming pipelines with orchestration, CI/CD and data quality tests.

    Platform layersPipelines & streaming

    Tools we use mostAirflow, dbt, Kafka, Spark

  2. Data Warehouse & Lakehouse

    Lakehouses and cloud warehouses, designed new or migrated off legacy systems.

    Platform layersLakehouse & warehouse, governance

    Tools we use mostDatabricks, Snowflake, Fabric, BigQuery

  3. Data Analytics & BI

    One set of metric definitions, a semantic layer and the dashboards built on it.

    Platform layersBI, governance

    Tools we use mostdbt, Power BI, Looker, Tableau

  4. AI & Machine Learning

    Retrieval assistants, agents and ML models, evaluated before release and traced in production.

    Platform layersAI & LLM, ML & LLMOps

    Tools we use mostOpenAI, Claude, Bedrock, LangGraph, MLflow

  5. Cloud Data Solutions

    Landing zones, infrastructure as code, access control and cost reporting.

    Platform layersCloud, infrastructure

    Tools we use mostAWS, Azure, Google Cloud, Terraform

  6. IT Outsourcing

    Engineers who join your team and its stand-ups, with our architects reviewing their work.

    Platform layersAny of the above

    Tools we use mostWhatever your team runs

The team on an engagement

A DATAI team is built from four roles. The written plan after the assessment says which ones your engagement needs and for how long.

  1. Architect

    Leads the assessment and decides the architecture. Reviews every merge request before it ships, whoever wrote it.

    What you get from them

    • The written plan: scope, team, timeline
    • Architecture diagram

  2. Data engineer

    Builds ingestion and pipelines in your repositories, with orchestration, CI/CD and schema, freshness and volume tests on each one.

    What you get from them

    • Pipeline code in your repository
    • Alerts routed to owners
    • A runbook per pipeline

  3. Analytics engineer

    Turns raw tables into dbt models and shared metric definitions, so revenue means the same thing in every report.

    What you get from them

    • dbt models and metrics
    • Semantic layer
    • Dashboards and reports

  4. ML / LLM engineer

    Builds retrieval assistants, agents and models on governed data, and keeps the evaluation set each release has to pass.

    What you get from them

    • Evaluation sets and LLMOps
    • Tracing of cost, latency and quality per request

Careers and the internship

We hire data, analytics and AI engineers for international projects. We also train new engineers through the Data & AI Engineering Internship, which is free and selective. Interns learn the same standard on real work.

See careers About the internship
1 November 202631 January 2027
This internship batch runs 92 days, drawn to scale by month.

Talk to a senior engineer

Your first call is 30 minutes with a senior engineer. Bring your stack and the decision you need to make; afterwards you get a short written summary of next steps.

Book a consultation Send a request