AI/ML Engineer - Data Pipeline

Role Overview :

You will build the data pipelines and diagnostic models for an AI platform that reads network and customer telemetry, finds the likely root cause of a broadband fault, and decides whether a technician visit is needed. You report to the ML lead and work in two week sprints.

What you will do :

- Build and run ingestion from ticketing, ACS, alarm, topology, DPI, billing and Kafka streams into one evidence record per ticket, with entity stitching from device to circuit to splitter to OLT port.

- Build and tune the diagnostic models : port the proven baseline scorer, add anomaly detection on time series, graph based root cause analysis on the network topology, and conformal uncertainty so the system knows when it is unsure.

- Score graded batches, tune thresholds against the client's recall and precision criteria, and keep the comparison against the legacy tool running in shadow mode.

- Wire the closed loop : callbacks and overrides captured as labels, weekly retraining, drift monitoring after go live.

- Write down every data quirk you find and explain model verdicts in plain language to L2 engineers and client platform owners.

What you bring :

- 2 or more years in data or machine learning engineering with at least one model you built running in production.

- Strong Python and SQL; exposure to a streaming platform such as Kafka; comfortable with containers and CI.

- Classification basics : feature pipelines, threshold selection, recall and precision trade offs.

- Strength in one of : data pipelines and data quality, or time series and graph based modelling.

- A measurement habit : a graded sample and a baseline before you believe a number.

- Clear written and spoken English.

Good to have :

- Graph neural networks or knowledge graphs.

- Conformal prediction or similar uncertainty methods.

- Large scale operational data such as tickets or alarms.