Kiavash Samadi

Software Development Engineer II · Data & ML Infrastructure

kiavashsamadi@gmail.com · github.com/kiavashsamadi · linkedin.com/in/kiavashsamadi

Software Development Engineer who ships production data and machine-learning systems at scale. Over three years at Amazon — promoted twice — I built ML and LLM-powered products, owned petabyte-scale data infrastructure, and led the reliability, performance, and cost work that keeps large systems healthy. I do my best work on ambiguous, cross-team problems that need someone to own them end to end.

Experience

Software Development Engineer

Amazon · Vancouver, BC

Intern → SDE I → SDE II · 2022 – Present

  • Designed and shipped a production LLM application on Amazon Bedrock that automatically assesses the cross-team impact of new projects, giving leadership early visibility into conflicts that used to surface only after launch.
  • Built and launched personalized recommendation models for Amazon Fresh and Whole Foods, owning the machine-learning pipeline end to end — from data to real-time serving.
  • Owned petabyte-scale data infrastructure behind grocery analytics and ML, cutting its cloud spend by over 90% while data volume kept growing.
  • Migrated petabytes of data into privacy-compliant (GDPR/CCPA) data lakes with zero downtime, and led a JDK 17 upgrade that significantly reduced compute cost across services.
  • Kept business-critical fulfillment services reliable at 99.9%+ availability across multiple AWS regions, leading incident response and on-call.
  • Re-architected scheduled ML workflows to run event-driven, cutting end-to-end run times from hours to minutes and removing manual operational toil.
  • Mentored engineers and interns and shaped technical direction through design and code review — promoted from intern to SDE II in three years.

Software Developer — Full-Stack & Security

Department of National Defence · Ottawa, ON

2019 – 2020

  • Built internal workflow applications with Spring Boot (Java) and Angular — e-signatures, notifications, and role-based access — improving end-to-end performance by 50% and serving as scrum master.
  • Stood up a full CI/CD platform (Azure, Docker, Kubernetes, Jenkins, Terraform, SonarQube) and authored an XML digital-signature library used across applications.
  • On the security side, contributed to Assemblyline (open-source malware analysis), building a Python service that de-obfuscates and emulates malicious Excel 4.0 macros to generate threat reports — without executing the payload.

Education

B.Sc., Computer Science

University of British Columbia

2017 – 2022

  • Coursework in algorithms, machine learning & data mining, NLP, databases, operating systems, and ethics.
  • CS Student Society — Social Officer & First-Year Representative.

Skills

Languages
Java · Python · TypeScript · SQL · Scala · Kotlin
Cloud / AWS
Lambda · Glue · S3 · DynamoDB · Step Functions · Kinesis · SageMaker · Bedrock · EventBridge · CDK
Data
Apache Spark · ETL · Parquet · data lakes · streaming · petabyte-scale pipelines
ML / AI
LLM apps on Bedrock (Claude) · recommendation systems · SageMaker · MLOps
Practices
Distributed systems · CI/CD · Docker · Kubernetes · privacy compliance · operational excellence / on-call