Telecom Descriptive Analytics Platform

I am building an AI-driven platform that helps telecom vendors understand exactly what has happened across their networks by turning raw data into clear, actionable insight. The focus is strictly descriptive analytics: I want to surface patterns, error trends, and usage anomalies from two key sources—real-time network performance metrics and our historical customer service logs—so engineers and account teams can quickly see where issues occurred and how widespread they were.

Here’s what I need from you:

• Data pipeline design and implementation that securely ingests high-volume performance counters alongside ticket and chat transcripts, cleans them, and stores them in a query-friendly format.
• Feature engineering and statistical/ML models (Python, pandas, scikit-learn or similar) tailored to summarising incidents, frequency, duration, and affected geographies or device types. Deep learning is optional if you have a compelling reason.
• A lightweight dashboard—Grafana, Kibana, or a React-based front end—that visualises daily, weekly, and monthly error trends, highlights top recurring issues, and allows drill-down to individual log entries.
• Documentation and a short hand-off session so my in-house team can maintain and extend the solution.

Acceptance criteria: data refresh completes within 15 minutes for each hourly batch; dashboards render key metrics in under two seconds; codebase is delivered via Git with clear README and unit tests.

If you have solid experience in telecom data, stream processing (Kafka/Spark/Flink), and building intuitive analytical UIs, let’s talk.

English level must be Native or C2