AI Data Team Lead

Алматы, улица Зеина Шашкина, 1/1

Метро: Абая

Описание вакансии

Location: Almaty, Kazakhstan Reports to: Director of AI R&D Team: Manages 3–4 analysts / BI specialists Level: Mid-manager (hands-on — this is a working manager role, not a pure people-management position) Languages: Russian required, English a plus

About the Role

We're looking for an AI Data Team Lead to own the analytical backbone of our AI systems — from the data that feeds our models to the dashboards that tell us whether they're working. You'll lead a small team while staying hands-on with SQL, Python, and Power BI, and you'll be the person who catches a data quality issue before it becomes a model problem, or a performance drop before it becomes a business problem.

This is not a modeling role — you won't be training or tuning models. But you'll need to understand how our models and AI agents work well enough to prepare the right data for them, monitor their behavior in production, and translate what you see in the numbers into decisions the business can act on.

We're specifically looking for someone proactive: someone who spots the ad-hoc analysis worth doing before it's requested, flags process improvements, surfaces insights nobody asked for but everybody needed, and doesn't wait for a pipeline to break before checking it.

Key Responsibilities

Team Leadership

Lead and mentor a team of 3–4 analysts/BI specialists supporting AI initiatives.
Set and maintain standards for Power BI development, datamart design, and reporting quality across the team.

AI Monitoring & Proactive Analytics

Design and maintain monitoring frameworks to track AI/agent performance: model health, automation rates, CSAT, FCR, and conversion impact.
Proactively run ad-hoc analyses to catch performance drift, data anomalies, or business shifts before they show up as complaints — don't wait to be asked.
Conduct root-cause analysis when AI performance deviates from benchmarks, distinguishing genuine model degradation from shifts in the underlying business context.
Proactively identify and propose process improvements across reporting, monitoring, and data workflows.

Data Engineering & Preparation

Own the preparation, quality, and labeling of datasets used for model training and evaluation by internal teams and external AI partners.
Perform data assurance on ETL pipelines built by IT/DWH — proactively check pipeline output quality rather than relying solely on IT sign-off.
Monitor pipeline health continuously; flag and drive resolution of data issues at the source.
Work with IT/DWH to ensure the data feeding AI systems is reliable, timely, and well-documented.

Strategic Cooperation

Act as the primary point of contact between AI business owners, IT/DWH, and (where relevant) external AI partners for data inputs and analytical outputs.
Translate business requirements (contact center, retail banking, etc.) into concrete technical data specifications.
Contribute your domain's practices and needs into the company's central Data Governance framework; ensure your team's work complies with it.

What Success Looks Like

Faster time-to-insight on AI/business questions.
Measurable contribution to automation rate increases (chatbot/voicebot), conversion improvements, and CSAT/FCR gains.
Early detection of data or model issues — before they surface as business complaints.

Required Skills & Qualifications

Leadership

Proven experience leading or senior-contributing within an analytics/BI team, with a track record of owning complex data deliverables.
Strong stakeholder management across business and technical (IT/DWH) teams.

Analytics & BI

Expert-level Power BI (DAX, workspace management, enterprise dashboarding).
Strong grasp of data modeling (star/snowflake schemas, datamart design).
Sharp analytical instincts — able to independently identify what's worth investigating, not just execute requested analyses.

Technical

Expert SQL across large-scale platforms (Oracle DWH, Impala/Hadoop).
Strong Python (Pandas/NumPy) for data prep and ad-hoc scripting.
Working knowledge of PySpark/Spark SQL for large transactional datasets.
Solid understanding of the AI/ML lifecycle (e.g. CRISP-DM) and how RAG architectures consume corporate data — enough to prepare data and monitor outcomes, not to build models.

Tools & Technologies