In January 2026, I was laid off from Amazon after almost 2 years as a Software Development Engineer. I had built distributed systems at scale, led cross-team initiatives, and earned top performance ratings. But none of that mattered when the layoff email arrived.
The layoff hit hard, but it also sparked an idea. Every time I applied for a job, I wished recruiters could just talk to my experience directly instead of scanning a 2-page PDF. What if my resume could actually answer questions?
That's how iku was born. I wanted to build a platform where any professional could upload their CV and instantly get a personal AI assistant that knows their career inside-out. No code, no setup complexity, just a shareable link.
I built the entire platform from scratch in 4 days: a serverless backend on AWS with 8 CDK stacks and ~195 resources, a Next.js frontend, RAG-powered chat with Bedrock and Pinecone, auth, analytics, document processing, and everything in between.
iku is also my portfolio. Every architectural decision, every line of code, every deployment pipeline represents what I can bring to your team. This page is my way of showing, not just telling.
8 CDK Stacks · ~195 AWS Resources
REST API with 21 endpoints, Cognito authorizer, request validation, and CORS configuration.
7 Python handlers: auth, chat, documents, billing, indexing worker, scheduler, and user management.
4 tables with GSIs for users, documents, chat sessions, and rate limiting. On-demand capacity.
RAG pipeline using Knowledge Bases, Titan embeddings (1024d), and Claude 3.5 Haiku for generation.
3 buckets for documents, photos, and frontend hosting. OAC-secured CloudFront distribution.
User pool with email verification, password policies, and JWT-based authentication flow.
Event-driven document indexing via queues. Scheduled jobs for ingestion safety net and cleanup.
This chatbot was trained on my resume and professional documents. Ask about my experience, skills, projects, or anything else.