About
Experienced Software Engineer specializing in the design and development of privacy-conscious, scalable AI and cloud-based systems. Proven leader in building robust, high-throughput solutions for critical functions like content moderation, identity management, and operational excellence. Leverages deep expertise in AWS, LLMs, and full-stack development to drive significant cost savings, enhance security, and improve developer efficiency.
Work
→
Summary
As an experienced Software Development Engineer at Amazon, Sky led the end-to-end design, development, and operation of critical, privacy-conscious, and scalable AI and cloud-based systems for Alexa, significantly enhancing security, efficiency, and compliance across various product domains.
Highlights
Architected and implemented pre-inference content moderation APIs on ECS Fargate, integrating ML models to annotate PII in LLM prompts and prevent privacy violations, scaling to support 500 TPS in LLM runtime.
Designed and deployed a near-real-time offline content moderation system for detecting privacy risks (PII, prompt injection, social engineering), integrating Mistral LLM to improve detection recall and enhance risk coverage.
Led end-to-end design and development of a multi-tenant user management system on ECS, providing AuthN/Z and SSO capabilities, which automated user pool provisioning and reduced enterprise onboarding time by 35%.
Engineered a system to classify over 110,000 Alexa Skills by analyzing developer attestations and privacy signals, identifying potential PII violations and establishing an allowlisting mechanism to improve precision in acceptable PII use.
Drove cross-team Operational Excellence initiatives, reducing CloudWatch costs by 95% via metric batching and S3 storage costs by 47% through tiered lifecycle optimization, with best practices adopted by 2+ teams.
Led full-stack development and operation of a Ruby on Rails web application for internal trainers, owning end-to-end delivery of key features including async job processing, report generation, and automated notifications.
Enhanced the risk-tiering data pipeline for Alexa Skill developers using ETL workflows across AWS Data Pipeline, S3, and DynamoDB, exposing insights via RESTful APIs to improve trust transparency.
Served as Scrum Master, actively driving sprint planning and improving team execution for multiple development teams.
Projects
RAG Chatbot for Alexa Developers
Hackathon Project
Summary
Retrieval-Augmented Generation (RAG) chatbot utilizing Amazon Q Business to answer questions from Alexa skill developers, connecting to and ingesting data from official Alexa developer documentation.
Skills
Programming Languages
Java, Ruby, Python, React, Typescript.
Cloud Platforms
AWS, API Gateway, Lambda, ECS, DynamoDB, RDS, S3, SNS, SQS, SES, Cognito, AWS Data Pipeline, AWS CDK.
Frameworks
Ruby on Rails, Spring, Django.
AI & LLM
OpenAI APIs, Retrieval-Augmented Generation (RAG), Prompt Engineering, AI Coding Agents (Cursor, Claude, Kiro), Mistral LLM.
Methodologies & Practices
Scrum, CI/CD, OS Patching, Security Compliance, Cost Optimization, Capacity Planning, ETL Workflows, Content Moderation, Identity & Access Management (AuthN, AuthZ, SSO), UAT, Cross-functional Collaboration, Agile Methodologies, System Design, API Design, Data Aggregation, Risk Mitigation, Full-Stack Development, Web Application Development, Report Generation, Automated Notifications, Async Job Processing.