Rajeev Paudel

I’m a data engineering team lead specializing in high-throughput, low-latency data infrastructure across blockchain, social, and media domains. I design petabyte-scale warehouses, real-time streaming pipelines, and in-warehouse AI/NLP workloads that turn raw events into decision-ready intelligence.

At 108 Capital, I lead a data engineering team and shipped five products from prototype to production, owning architecture, delivery, and reliability. I delivered a multi-chain EVM on-chain indexer to production across five networks, managing 20TB+ of warehouse data with real-time streaming through Apache Kafka and on-the-fly decoding via ClickHouse UDFs.

I also run NLP model UDFs on a distributed, GPU-enabled ClickHouse cluster for sentiment analysis and entity extraction directly at query time. Beyond indexing, I built an AI transcription pipeline that auto-discovers YouTube videos and podcast episodes per channel, orchestrated with Ploomber and Celery for scalable background processing.

Previously, I built a proprietary on-chain data indexer from scratch with ClickHouse and authored the ETL using dbt and Prefect, achieving sub-1-minute latency from on-chain events to the warehouse. I’ve also engineered large-scale social ingestion and led migration from ClickHouse to Trino while benchmarking cluster configurations and initiating Spark-based ETL direction.