You are tasked with designing a data processing pipeline using Apache Beam that will receive streaming data from Cloud Pub/Sub and enrich this data with static reference data stored in BigQuery. It is important to note that the reference data set is small enough to fit entirely in memory of a single worker machine. The final output of the pipeline, which includes the enriched data, should be written back to BigQuery for further analysis. Given these requirements, what type of job and specific transforms should you utilize in this pipeline?