Radha Garine – Senior Software Engineer

About

A seasoned software engineer with a career spanning distributed systems, real-time streaming, AI-native platforms, and embedded software across telecom, finance, retail, automotive, and enterprise domains. I have worked across the full stack — from low-level C++ on embedded devices to cloud-scale Kafka pipelines and GPT-powered agent systems — at companies like Verizon, Amazon, Microsoft, and John Deere.

My current focus is on real-time streaming (Kafka Streams, RocksDB) and building AI-native automation that is reliable, auditable, and designed to scale under production pressure.

50M+Events processed / day

115MDAU on Microsoft Teams

90%Performance gain at Oracle

$B+Cash flows reconciled (Amazon)

3FAANG-tier companies

Technical Skills

Languages

C++17/14/11Java 8/11/17PythonC#TypeScriptJavaScriptPerlShell ScriptC

Cloud & Platforms

AWSAzureLinuxWindowsDockerKubernetesNode.js

Databases

OracleMySQLInformixPostgreSQLMongoDBCosmos DBDynamoDBCassandraKusto

Streaming & Infra

Kafka StreamsRocksDBSpring BootSpring MVCCI/CDGitNewRelicSplunk

AI & ML

OpenAI GPTNLPRAGLangChainAI AgentsLiteLLMSupabaseNeural Networks

Architecture

MicroservicesDistributed SystemsMVCClient-ServerEmbedded / HIDTCP/IP SocketsSiebel CRM

Work History

VerizonAug 2025 – Present

Senior Software Engineer

📍 Irving, TX

Built a real-time Kafka Streams application (Java Spring Boot) processing 50M+ events/day for mission-critical telecom billing.
Eliminated consumer lag by fixing a misconfigured RocksDB global memory limit; reduced disk writes from 2.7 GB/s → <260 MB/s via Universal Compaction and Bloom Filters.
Scaled system to process 40 Data Groups across 6 servers with healthy load averages.
Architected the RBM Production Support Automation Platform — an enterprise AI Agent system (Python, FastAPI, LiteLLM) automating Oracle billing remediation with PII redaction and durable checkpointing.
Designed Agent vs Worker pattern with FlowDefinition extensibility registry; enforced safe Oracle writes via centralised sql_executor layer.

Appsboat Inc.Sep 2024 – Jul 2025

Full Stack & AI Engineer

📍 Bellevue, WA

Designed a microservices-based AI SaaS platform (Python, C++, TypeScript, Next.js, React, Supabase, OpenAI) serving multi-tenant real-time analytics and conversational AI.
Built real-time analytics dashboard with BI insights using Recharts; implemented secure auth and state management with Zustand.
Integrated GPT-powered NLP/RAG agents for context-aware, knowledge-base-driven customer interactions.

AmazonJun 2021 – Jul 2024

Software Development Engineer II

📍 Seattle, WA

Spearheaded Checkbook — Amazon's automated bank reconciliation system handling thousands of monthly reconciliations and billions in cash movements for JPMC and HSBC.
Led the Scotia Migration Project: migrated Payout Pipeline to a microservice architecture on PayDock with more secure transmission protocols.
Contributed to the RTP Eligibility Project: built plugin infrastructure to streamline Real-Time Payments eligibility checks via the Payments Method Authority (PMA).
Stack: Java 8/11/17, Spring MVC, Maven/Gradle, AWS, Git, CI/CD.

MicrosoftJan 2020 – May 2021

Senior Software Engineer

📍 Bellevue, WA

Designed device integration features in Microsoft Teams using Modern C++ (C++17), serving ~115M daily active users.
Built HID components on Windows (WDK) and macOS (IOKit) for real-time hardware call controls (answer/hang-up, mute/unmute).
Integrated Azure services; contributed to system-wide performance tuning and debugging.

Neiman MarcusMar 2018 – Jan 2020

Lead Software Engineer

📍 Irving, TX

Led a team of ~10 engineers; drove architecture, design, and development of the mobile Point-of-Sale (MPOS) solution deployed across Neiman Marcus stores.
Designed Windows Services and Web APIs for MPOS; developed TCP Socket-based communication between MPOS and backend servers.
Delivered Customer Master Data (CMD) solution to centralise customer data on cloud; handled NMD Returns for POS.

VerizonJun 2016 – Mar 2018

Senior Software Engineer

📍 Irving, TX

Designed and developed Just-In-Time Billing systems using C++, Python, MySQL, and Oracle for millions of enterprise telecom records.
Built custom C++ plug-ins and stored procedures to handle high-volume usage data; led production monitoring strategy.

Key Projects

JavaCEA — Real-Time Data Pipeline StabilizationVerizon · 2025–2026

Transformed the JavaCEA Kafka Consumer into a stable, high-throughput real-time event aggregation system. Eliminated consumer lag via indexed RocksDB lookups, resolved "death spiral" crashes from a 10KB global memory cap, and reduced disk I/O from 2.7 GB/s → <260 MB/s using Universal Compaction and Bloom Filters. Scaled to 40 Data Groups across 6 servers.

Kafka StreamsRocksDBJava Spring BootCassandraNewRelic

RBM Production Support Automation PlatformVerizon · Apr 2026

Architected an enterprise-grade AI Agent system automating Oracle-based billing remediation. Agent vs Worker pattern with FlowDefinition extensibility registry; centralised sql_executor for safe DB writes; mandatory PII redaction; durable checkpointing (AsyncVegasSaver); FastAPI endpoints; RAG/runbook hooks.

AI AgentsPythonFastAPIOpenAI / LiteLLMOracle DBRAG

Amazon Checkbook — Bank Reconciliation at ScaleAmazon · 2021–2024

Spearheaded development of Checkbook, Amazon's automated bank account reconciliation system handling thousands of monthly reconciliations and billions in cash movements. Designed remittance resource generation for major banking partners JPMC and HSBC, enabling seamless ledger line matching (USL).

JavaSpring MVCAWSMicroservices