Close

Susindaran Elangovan

Software Engineer

Distributed Systems · Search · Data Infrastructure

Download Resume

About Me

I build distributed systems, search platforms, and data infrastructure that teams can depend on at scale. At Stripe, I lead cross-company architecture and execution for mission-critical platforms, from unifying Dashboard search to migrating billions of documents and improving ingestion across hundreds of terabytes of data. My work has delivered multi-million-dollar annual savings, substantial latency reductions, and reusable infrastructure adopted across teams.

I focus on turning complex infrastructure problems into shared platforms: aligning teams on architecture, planning phased rollouts, and building recovery, authorization, and operational reliability into the design. I also mentor engineers and build tools that make on-call analysis reliable and repeatable.

Experience

Stripe

Software Engineer, Search & Data Infrastructure

  • Architected and drove a cross-company overhaul of Dashboard list-view serving, replacing fragmented backends with a unified event-driven search platform. Coordinated enrichment, massive backfills, and phased rollout, cutting organization-wide load times from 15 seconds to 5 seconds and establishing a broadly adopted architectural standard.
  • Designed a reusable migration platform to repartition 26 billion documents across 32 TB, with parallel execution, checkpointed recovery, and live dual writes. Enabled large-scale infrastructure upgrades without product-code changes.
  • Rearchitected search ingestion for 450 TB across five Elasticsearch clusters, saving $3 million annually, tripling per-processor throughput, and reducing p99.9 ingestion lag by 90% to under 12 seconds.
  • Used CPU profiling and query analysis to eliminate 235 million database reads per day, reduce p99 hydration latency by 90%, and save $291,000 annually.
  • Delivered secure cross-user caching for Trino-backed analytics, returning cache hits in approximately 100 ms while preserving fine-grained authorization and enabling scalable live-query dashboards.
  • Published two internal AI agent plugins for Trino query triage and cluster scaling reports. Used deterministic AST-based configuration parsing to support reliable, repeatable on-call analysis across the team.
  • Implemented India data residency through regional ingestion, cross-cluster search, and mTLS while maintaining availability. Partnered with Security and IAM to mentor an intern on context-aware search authorization.

Evernote

Software Engineer Intern

  • Designed and developed a Java gRPC microservice for note-content data in a dynamically sharded MySQL store, and modified the Java backend to use multi-version concurrency control for distributed transactions.
  • Identified a hot table and executed a zero-downtime migration to a separate datastore, reducing load on a single-point-of-failure datastore by approximately 15%.

Flipkart

Software Development Engineer

  • Core member of the team that designed and built a configurable, fault-tolerant accounting service to validate, enrich, and audit organization-wide accounting events at low latency.
  • Designed and built TaaS, a RESTful service for asynchronous test execution and automated email reporting.

Education

M.S. Computer Science - Intelligent Systems

The University of Texas at Dallas

2019

B.E., Computer Science and Engineering

Anna University, India

2015

Selected Infrastructure Work

A shared search architecture

Unified fragmented Dashboard list-view backends around an event-driven search platform. The work combined cross-company architectural alignment, data enrichment, large backfills, and a phased rollout into a shared standard adopted across teams.

Impact: Organization-wide list-view load times fell from 15 seconds to 5 seconds.

Reusable migration infrastructure

Built a platform for repartitioning large search datasets with parallel execution, checkpointed recovery, and live dual writes. The reusable design enabled infrastructure upgrades without changes to product code.

Scale: 26 billion documents across 32 TB.

Faster, more efficient search ingestion

Rearchitected ingestion across five Elasticsearch clusters to improve throughput, reduce tail lag, and lower operating costs for a 450 TB search footprint.

Impact: $3 million in annual savings, 3× per-processor throughput, and p99.9 ingestion lag below 12 seconds.

Skills & Technical Focus

Architecture & Technical Leadership

Distributed Systems

Data & Infrastructure

Languages

Get in Touch