Price Prediction API
Using ML for automated price estimation
Project context: External blogpost
Personal contributions:
- Led development of 3 backend APIs serving ML models for real estate price prediction, leading to extended client engagement with 2 more APIs, by following
MLOpsbest practices andCI/CDwith Kubernetes; - Owned the development of 5 ML models for real estate pricing, while also serving as primary on-call for the 5 APIs on
KubernetesCI/CD, operating across different markets for an international client; - Reduced recurring GCP storage costs by optimizing ML output file size by an order of 10;
- Reduced latency per ML API call by 3s, by caching data for queries on GCP
BigQuery; - Configured access for team members with
GCPIAM and K8s RBAC; - Handled data ingestion, process and monitoring
data pipelinesfor weekly model retraining;