Yashaswini Chandrappa
Software Development Engineer at Amazon | Distributed Systems & LLM Evaluation Pipelines | AWS, Python, Java | Building AI-powered systems at scale
- Role
- Software Development Engineer at Amazon
- Location
- Seattle, WA, US
- LinkedIn followers
- 500 followers
About Yashaswini Chandrappa
Software Engineer with 5+ years of experience designing and scaling distributed systems at Amazon, delivering production systems serving millions of users.Currently building LLM-based evaluation systems for Amazon\'s Seller Assistant AI, processing 500K+ daily requests to ensure quality across 750K+ sellers. Previously built critical pharmacy infrastructure including a digital shipment literature service that reduced paper usage by 95%+ and real-time labor tracking pipelines deployed across 14 fulfillment centers.What I do best:→ Design and implement distributed systems and event-driven architectures→ Build LLM-powered evaluation and quality control pipelines→ Drive operational excellence and mentor engineers→ Deliver end-to-end solutions from design to productionCore Technologies- Languages: Java, Python, JavaScript/TypeScript, Ruby, Go- AWS: Lambda, Step Functions, Bedrock, DynamoDB, S3, SNS/SQS, API Gateway, SageMaker, EventBridge, CDK- Systems: Distributed Systems, Microservices, Event-Driven Architecture, REST APIs, State Machines- Frontend: React.js, AngularJS, Vue.jsI\'m passionate about building reliable, scalable systems and always looking to connect with fellow engineers and tech enthusiasts. Feel free to reach out!
Experience
Software Development Engineer
Mar 2025 — Present · Seattle, WA, US
Building LLM-based evaluation systems to ensure quality and reliability of Amazon\'s AI-powered Seller Assistant serving 750K+ sellers.Key Contributions- Architected auto-evaluation system using AWS Bedrock, Step Functions, and Langfuse, scaling from 300 manual audits/week to conversations/day (~500K LLM calls/day)- Designed two-phase quality control feedback loop using Lambda, SageMaker Ground Truth, and Step Functions with configurable disagreement thresholds triggering automated re-audits- Implemented end-to-end CDK infrastructure pipeline with secure cross-service authentication using transitive auth tokens- Led operational excellence initiatives including alarm aggregation and automated report generation, improving production support efficiencyTechnologies: AWS Bedrock, Step Functions, Lambda, SageMaker Ground Truth, DynamoDB, S3, CDK, Python, Langfuse
Education
Northeastern University
Master of Science - MS, Computer science and Information systems
2019 — 2021
Visvesvaraya Technological University
Bachelor of Engineering (B.E.), Computer Science
2012 — 2016
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.