Case Study: AlphaSense cuts reindexing from one week to 2 hours and slashes costs with Amazon Web Services

AlphaSense - Customer Case Study

AlphaSense is a San Francisco–based fintech startup that offers a semantic search engine to surface investment insights from hundreds of millions of documents. As its dataset grew, re‑indexing for deep natural‑language analysis became increasingly slow and expensive—taking a week or more and becoming cost‑prohibitive for timely, granular search results.

AlphaSense moved its platform to AWS, using EC2/EBS, S3, RDS, EMR and—critically—EC2 Spot Instances with a parallelized, fault‑tolerant HPC architecture and optimized bidding. By running ~500 Spot Instances and tuning for restartability and security (IAM, VPC, encryption), it cut full reprocessing from a week to under 2 hours and reduced cost from about $5,000 to $80, while freeing engineers, enabling real‑time updates, and improving flexibility for product innovation.

AlphaSense

Raj Neervannan

Co-founder and CTO