Krithin Sitaram

Summary

Senior full-stack engineer specializing in geospatial data systems for transit planning, with 12+ years of industry experience including infrastructure work at Google. Builds geospatial pipelines end-to-end — OSM/POI extraction, PostGIS analytics, H3 clustering, and isochrone routing — and owns features from Postgres/PostGIS through React/Redux map UI to production. Owned epics for ride-request recommendations and demand-flow automation; primary code reviewer (~1,550 PRs) across backend, routing engine, and map client.

Experience

• Owned the ride-request recommendations epic: designed and shipped the backend for zone recommendations with H3 hex clustering, a suggestion API, and per-subservice PostGIS query optimization.
• Built OSM POI zone insights end-to-end (Rails/PostGIS pipelines with category weighting and a React insights sidebar), and authored the tech design and v1 auto-sync pipeline that provisions demand-flow geo datasets for newly live Via cities — replacing manual dataset ops.

• Productionized Jane drive-to-transit access for RTA Chicago (place types, map UI, walkshed/frequency/network isochrone modes) and migrated driving-time computation from Google Maps to self-hosted Valhalla — batched sources_to_targets routing, perf instrumentation, infra hardening; scaled Jane workers 2×.
• Co-inventor on a patent covering accessibility-analysis methods developed for the Jane engine (U.S. Patent No. [Number / “Pending”]).

• Shipped self-serve geo dataset upload (zip/CSV ingestion, translated error handling, extraction diagnostics) and geodata foundations — layer categories, CSV column autodetection, PostGIS schema and index tuning; built the NY Bus Speeds dataset pipeline (extractor, tileserver, client rendering).
• Architected a continent-wide OpenStreetMap extraction pipeline in Go that generates tiled pedestrian-accessibility road-network data for isochrone analysis; built the data-ingestion layer of an Apache Airflow ETL processing location data from hundreds of thousands of shared e-scooters across 15 operators in major US cities — designing schema, tuning indexes, and driving data-quality investigations with external partners.
• Leading an in-progress migration of data-science models (random forest, ridge regression) off SageMaker/MLflow onto custom ECS services — replacing a manual, half-day-per-model deploy (two new SageMaker endpoints each) with a partially self-serve workflow that supports multiple model deployments per day.

• Tools & Infrastructure (SETI): helped launch a near-real-time, log-based verification service checking 200+ business-logic rules across teams; built a cache layer for low-latency streaming database joins scaling to ~10⁶ log entries/sec, plus production monitoring and alerting.
• Site Reliability Engineer: simplified configuration management for Google’s global load balancer, breaking a monolithic production push into smaller, more reliable steps with richer ACLs.

Education