✕ About Skills Experience Projects Education Contact
Available for Opportunities

Varshitha Nandyala

Backend & Distributed Systems Engineer with 5+ years building fault-tolerant, horizontally scalable systems — from fintech payment rails to production AI infrastructure.

Get in Touch → View My Work
⚡
10M+
Transactions/day at peak load
🛡️
99.99%
System uptime achieved
💡
5TB+
Data processed daily via ETL
🤖
−30%
LLM hallucination rate reduction
About Me

Building systems that don't flinch under pressure

I'm a backend and distributed systems engineer who thrives at the intersection of scale, reliability, and real-world impact. At Razorpay, I designed payment microservices processing millions of transactions daily; at Snowflake, I've engineered multi-TB ETL pipelines and built Agentic AI evaluation infrastructure.

My work lives in the details: idempotency keys that prevent double-charges, circuit breakers that degrade gracefully under load, and partition-aware Kafka topologies that never lose a message. I write systems you can trust at 3 AM.

Currently pursuing my Master's in Information Technology at Auburn University Montgomery, deepening my foundations while applying them at the frontier of production AI.

🏗️ Distributed Systems
🔄 Event-Driven Architecture
☁️ Multi-Cloud (AWS/GCP/Azure)
🤖 AI/LLM Infrastructure
🔒 HIPAA / PCI-DSS
📊 Observability

// Career Highlights

Fraud prevented (annual)₹50Cr+
Settlement discrepancies reduced−45%
Monthly cloud costs saved$15K
Manual QA cycles eliminated−70%
Release failures reduced−35%
Decision-making speed+60%

// Stack Snapshot

Primary LanguagesGo · Python · SQL
MessagingKafka · RabbitMQ
CloudAWS · GCP · Azure
OrchestrationK8s · Helm · Terraform
AI ToolingLangGraph · MCP · n8n
Technical Expertise

What I build with

🔧
Languages & Frameworks
Golang Python FastAPI Django REST TypeScript Node.js SQL Shell
🌐
Distributed Systems
Fault-Tolerant Architecture Idempotency Circuit Breakers Exactly-Once Delivery Horizontal Scaling Graceful Degradation
📨
Messaging & Streaming
Kafka RabbitMQ Consumer Groups Dead-Letter Queues Event-Driven Architecture Async Pipelines
☁️
Cloud & DevOps
AWS (EKS, Lambda, IAM) GCP Azure (AKS) Kubernetes Terraform Docker GitLab CI/CD Helm
🤖
AI & LLM Engineering
LangGraph MCP n8n Snowpark ML LLM Evaluation Agentic Orchestration Prompt Engineering
📊
Data & Observability
Snowflake PostgreSQL Redis MongoDB Prometheus Grafana ELK Stack ETL/ELT
🔒
Security & Compliance
HIPAA PCI-DSS mTLS Field-Level Encryption RBAC Tokenization Audit Logging
🔌
APIs & Backend
REST APIs GraphQL WebSockets OAuth 2.0 JWT Microservices High-Throughput Async
Work History

Where I've built

Aug 2024 – Present Remote, USA
Software Engineer
Snowflake · Remote, USA
  • Designed and scaled distributed Python-based ETL pipelines (Snowpark, Pandas, SQL) processing 5TB+ data daily across partitioned workloads, reducing execution time by 40% through query optimization and parallel task execution.
  • Engineered Snowflake multi-cluster distributed compute architectures with workload isolation and auto-scaling; applied advanced SQL tuning, result caching, and materialized views to cut monthly costs by $15K.
  • Architected a distributed Agentic AI evaluation platform using LangGraph state-machine orchestration and MCP-based workflow automation (n8n); fault-tolerant multi-node execution across 5+ model variants reduced hallucination rates by 30% and cut manual QA cycles by 70%.
  • Developed HIPAA-compliant data governance frameworks (dynamic masking, row-level security), securing 1M+ sensitive records.
  • Deployed ML pipelines using Snowpark ML improving fraud detection accuracy by 22% in production; automated CI/CD pipelines using GitLab CI/CD, Docker, and Terraform reducing release failures by 35%.
PythonSnowflakeSnowparkLangGraphMCPn8nAWS LambdaHIPAATerraform
Jun 2020 – Jul 2023 India
Software Engineer
Razorpay · India
  • Built high-throughput async payment microservices in Golang handling 10M+ transactions/day across UPI, cards, and wallets; enforced distributed consistency using idempotency keys, exponential-backoff retry, and circuit-breaker patterns, achieving 99.95% transaction reliability.
  • Architected distributed event-driven payment reconciliation using Kafka consumer groups and RabbitMQ dead-letter queues with exactly-once delivery semantics; reduced settlement discrepancies by 45% and increased authorization rates by 18%.
  • Developed a distributed real-time fraud detection platform with horizontally scalable rule evaluation; prevented ₹50Cr+ in fraudulent transactions annually with a false-positive rate below 0.1%; implemented PCI-DSS compliant tokenization for 500K+ merchants.
  • Designed multi-cloud infrastructure across AWS, Azure, and GCP using Kubernetes (EKS, AKS, GKE) with cross-cluster networking, maintaining 99.99% uptime for payment systems.
  • Built observability solutions using Prometheus, Grafana, and ELK Stack; integrated Jenkins, Docker, Kubernetes and canary deployments, reducing payment failure MTTR by 50%.
GolangKafkaRabbitMQKubernetesAWSGCPAzurePCI-DSSPrometheusGrafana
Featured Work

Projects & Systems Built

🔐
Razorpay
Distributed Fraud Detection Platform
Horizontally scalable real-time fraud detection with ML model outputs integrated into configurable velocity checks and stateful risk scoring across distributed nodes. PCI-DSS compliant with field-level encryption and RBAC for 500K+ merchants.
₹50Cr+ fraud prevented annually · <0.1% false-positive rate
GolangML IntegrationPCI-DSSRedisRBAC
💳
Razorpay
Payment Reconciliation Engine
Partition-aware event-driven reconciliation system using Kafka consumer groups and RabbitMQ DLQs with exactly-once delivery semantics. Dynamic routing algorithms for acquirer selection across distributed nodes.
Settlement discrepancies −45% · Authorization rates +18%
GolangKafkaRabbitMQExactly-Once
🤖
Snowflake
Agentic AI Evaluation Infrastructure
Fault-tolerant, parallelized multi-node LLM evaluation platform using LangGraph state-machine orchestration and MCP-based workflow automation. Processes 10K+ test cases per run across 5+ model variants for tool-calling accuracy, reasoning consistency, and response determinism.
Hallucination rates −30% · Manual QA cycles −70%
LangGraphMCPn8nPythonMulti-node
🗄️
Snowflake
Distributed ETL & Analytics Pipeline
Python-based ETL pipelines processing 5TB+ daily across partitioned workloads with advanced query optimization, parallel task execution, and partition pruning. Real-time analytics spanning AWS Lambda, Python microservices, and Snowflake for sub-second latency.
Execution time −40% · Decision-making speed +60%
SnowparkPythonAWS LambdaMaterialized Views
🏥
Snowflake
HIPAA Data Governance Framework
Comprehensive HIPAA-compliant data governance with dynamic masking, row-level security, and audit trails securing 1M+ sensitive health records. Integrated CI/CD with zero-downtime deployment strategy.
1M+ sensitive records secured · Zero-downtime deployments
HIPAASnowflakeRow-Level SecurityTerraform
☁️
Razorpay
Multi-Cloud Payment Infrastructure
Geo-redundant multi-cloud infrastructure across AWS, Azure, and GCP using Kubernetes (EKS, AKS, GKE) with cross-cluster networking, Lambda-based event processing, replication, and failover strategies for mission-critical payment workloads.
99.99% uptime · 10M+ daily transactions
AWSGCPAzureKubernetesGeo-Redundancy
Education

Academic Background

🎓
Auburn University at Montgomery
Master of Science in Information Technology
Aug 2023 – May 2025 · Montgomery, Alabama
Deepening expertise in distributed systems, software engineering, and AI systems while applying academic foundations to production engineering challenges at Snowflake. Strong focus on scalable architectures, data platforms, and the evolving landscape of intelligent systems.
Let's Work Together

Ready to build something that scales?

I'm open to backend engineering roles, distributed systems challenges, and AI infrastructure opportunities. If you're building something ambitious, let's talk.

✉ varshithan0912@gmail.com