Senior Full Stack Engineer with 8+ years of experience building scalable web applications and cloud infrastructure for remote teams across the US, Europe, and Asia. Strong backends in Node.js and Ruby on Rails, fast frontends in React, Next.js, and TypeScript. Owns DevOps end to end on AWS, GCP, and Oracle Cloud.
Spent the last two years deep in AI engineering multi-agent benchmarks, LLM evaluation, training data pipelines, production RAG systems, and agentic workflows. Previously engaged with Amazon Web Services as an LLM Training and Data Engineer. Currently building AI infrastructure at Turing. Mentored 50+ developers into full stack roles worldwide.
Frontend: React, Next.js, TypeScript, Tailwind CSS, Redux, Vue.js, HTML5, CSS3, Responsive Design
Backend: Node.js, Express.js, Ruby on Rails, Python, FastAPI, REST APIs, GraphQL, Microservices
AI and Agents: Multi-Agent Systems, LLM Benchmarking (SwarmBench, Harbor), Agentic Pipelines, Prompt Engineering, RAG Pipelines, Vector Databases (ChromaDB), Embeddings (sentence-transformers), RLHF-style Data Verification
DevOps and Cloud: AWS (EC2, S3, Lambda, API Gateway), GCP, Oracle Cloud, Docker, Kubernetes, GitHub Actions, CI/CD
Databases: PostgreSQL, MongoDB, MySQL
Testing: Jest, RSpec, PyTest, Cypress, Playwright
CMS and E-commerce: Shopify, WordPress, Webflow, Squarespace, Strapi, Framer
Customer support agent built with Amazon Bedrock, Amazon Nova Micro model, TypeScript, and AWS SDK v3. Demonstrates multi-turn conversation memory and agentic patterns using the Converse API. View Repository
Self-hosted multi-agent platform running 24/7 on Oracle Cloud. Three agents handle job discovery and scoring, resume tailoring and email applications, and LinkedIn alert parsing. Fetches live jobs from 7 APIs, scores them with Gemini Flash LLM, and delivers daily reports via Telegram bot and email. Built with Python, LLM orchestration, cron, and systemd.
Production RAG pipeline for document question answering. Uses sentence-transformers for embeddings, ChromaDB as the local vector database, and Gemini Flash as the answering model. Deployed on Oracle Cloud with a Telegram bot interface running 24/7 via systemd. Zero cloud spend.
Marketing site for an AI product. Built with TypeScript, Next.js, and Framer Motion. www.piace.ai
Arts organization site with headless CMS (Payload) and media pipeline (Cloudinary). Built with Next.js 15, Tailwind v4, Framer Motion, and Lenis. chicagoindianarts.org
Improved dataset preparation efficiency by 30% through Python automation during Amazon Web Services engagement.
Mentored 50+ junior developers into full stack roles over five years at Microverse.
Delivered 25% site performance and traffic improvements through optimization across multiple client projects.
