About this role
About the company Tabby builds financial products used by millions of users across the GCC. The infrastructure behind them runs at scale, under strict requirements for reliability, cost efficiency and regulatory compliance. This is not a course and not a shadowing programme. It is an engineering role with real responsibility. Context The Data Platform team runs the infrastructure that AI, ML and data workloads at Tabby depend on: compute, orchestration, deployment, observability and cost control across cloud environments. The work sits between classic DevOps and the machine-learning side. The same clusters, pipelines and monitoring that keep a service alive also keep models trained, served and measured. The internship is designed for strong early-career engineers who are comfortable in Linux and a cloud, and who already use AI tools in their own work rather than reading about them. Interns join the team, work on real production infrastructure under senior review and are expected to meet engineering standards from day one. What you will do This is not a helper or ticket-closing role. Interns work on real production tasks under senior review. Work with the cloud infrastructure (primarily GCP ) and the bare-metal fleet that host our data, ML and AI workloads Build and maintain CI/CD pipelines for services and models Run and troubleshoot containerised workloads on Kubernetes Set up and improve monitoring, alerting and logging, and act on what they show Automate repetitive operational work with Python or Bash instead of repeating it Support model training and inference workloads: environments, resources, deployment, cost Investigate incidents in infrastructure and pipelines and help find root causes Improve the reliability and cost efficiency of the platform Work within SAMA regulatory requirements : in Saudi fintech, where data lives and who can reach it is part of the engineering problem, not paperwork someone else handles What you will actually work with Not a wish list. This is the stack the team runs today. Nobody is expected to arrive knowing all of it. Data: CDC pipelines, BigQuery, Airflow ML: Airflow, ClearML and similar orchestration and experiment tooling AI: bare-metal GPU servers, vLLM, open-source models served in-house Platform: GCP, Kubernetes, Linux, networking Context: SAMA regulations