
Ramneeq S Soddhi
Software Development Engineer II | Amazon
Hire this person
Send a job offer directly to this candidate
About
Software Development Professional with over 4 years of experience in Back-end Engineering, Full-Stack Development, Distributed Systems, Cloud-Native Applications, Generative AI, LLM Integration, and Microservices Development. Currently working as Software Development Engineer II at Amazon, developing AI-powered engineering tools, modernizing back-end platforms, and strengthening production reliability across customer-contact processing systems. Architected reusable checkout UI frameworks at Amazon Autos, consolidating components across 30+ pages and supporting customer-facing experiences serving millions of customers quarterly.
Distinctive experience in engineering reusable autonomous solutions, including PipelinePlumber, an AI agent diagnosing failures across 50+ CI/CD pipelines, validating fixes through sandbox builds, and generating change requests, alongside CloudWatch-based log centralization reducing discovery time from 30+ minutes to few minutes.
Experience
Software Development Engineer II | CS Proton – Back-end Platform & AI Tooling at AMAZON (2022-10 – Present)
- LLM-Powered Email Automation: Independently designed and delivered two LLM-based email automation workflows for Account Closure and Abuse Detection, addressing an annual opportunity of approximately 1M customer contacts (3% of total) across both use cases, comprising ~700K contacts for account closure and ~300K contacts for abuse detection, within a platform processing 33M annual contacts and achieving an overall automation rate of 14% (~4.5M contacts).
- Autonomous AI Agent Development: Designed and built PipelinePlumber, a reusable AI agent diagnosing CI/CD failures across 50+ engineering pipelines, validating proposed fixes through sandbox builds, and generating change requests to streamline failure investigation and support adoption across additional engineering pipelines.
- Production Debugging Modernization: Migrated legacy email-service logs to AWS CloudWatch and enabling centralized log queries with LLM-assisted production debugging, reducing log-discovery time from 30+ minutes to a matter of minutes.
- Back-End Platform Modernization: Modernized core contact-processing services by transitioning away from a deprecated framework and developing Java APIs, clients, authentication mechanisms, metrics, and alarms while maintaining backward compatibility.
- Production Incident Recovery: Restored LLM-based email classification services following a complete outage by tracing the failure to a retired AWS Bedrock model and deploying a supported replacement to reinstate service availability.
- Engineering Team Development: Conducted technical interviews for engineering recruitment and mentoring an intern through a six-month development period, supporting their transition into a full-time engineering position.
Software Development Engineer I | Amazon Autos & CS Proton – Checkout / Email Automation at Amazon
- Checkout UI Architecture: Established ownership of Amazon Autos checkout UI as the primary point of contact, developing customer-facing pages across desktop, mobile web, and the Amazon application while centralizing the shared UI widget framework to reduce integration effort by 65%+.
- Event-Driven Notification Architecture: Designed and implemented a notification service from the ground up using AWS SNS, SQS FIFO, and Lambda to orchestrate customer notifications for financing decisions, document signing, and order cancellations.
- Dynamic Checkout Experience: Integrated dynamic content widgets across 30+ checkout pages, accelerating offer updates from 1–2 days to under 15 minutes and optimizing front-end asset testing cycles from 90+ minutes to under 10 minutes.
- Generative AI Security Architecture: Implemented AWS Bedrock Guardrails through AWS CDK to protect 18+ categories of sensitive data and introducing prompt-injection defenses within customer-service email automation workflows.
- LLM Prompt Evaluation Framework: Designed and developed an LLM prompt-testing API incorporating 100+ regression scenarios to validate prompt modifications and minimize regression risks across customer-service automation workflows.
Education
Bachelor of Technology (B.Tech.) – Netaji Subhas Institute of Technology (NSIT), New Delhi (2018 – 2022)

