Skills & Projects
Selected builds across frontend, backend, product engineering, and applied AI—paired with the tools used to ship them.
Projects
DocuFlow AI
Serverless, event-driven platform on AWS that securely processes PDF, JPG and PNG invoices and receipts. Amazon Textract extracts the financial data, an AI proxy normalizes it, and low-confidence documents are routed through a human review workflow.
- React 19
- TypeScript
- AWS SAM
- Amazon S3
- AWS Lambda
- +6 more
VALSEA Classroom Copilot
Speech-first classroom copilot for Vietnamese–English code-switching lectures. Realtime mic → VALSEA ASR → structured notes, key terms, English recap, quizzes, exports and session history.
- Next.js
- TypeScript
- Fastify
- WebSocket
- Realtime ASR
- +2 more
TAIKHOANXIN
A digital access marketplace offering affordable licensed accounts through fast purchasing flows, streamlined product discovery, and dedicated customer-facing and admin experiences.
- Next.js 14
- TypeScript
- Tailwind CSS
- Ant Design
- SCSS
- +4 more
LUXE WEAR AI
Multi-tenant SaaS for building, deploying and monitoring AI agents. Agents answer from a tenant’s own knowledge base through RAG, can be embedded as a chat widget, and are tracked with usage and analytics dashboards.
- Next.js 14
- TypeScript
- Gemini AI
- Tailwind CSS
- Vector DB
- +6 more
DevOrbit - Blog Platform
An open-source blog platform for developers: Markdown articles with syntax highlighting, SEO optimization, dark mode, GitHub-powered comments and a reading experience built around typography.
- Next.js 16
- React 19
- TypeScript
- Tailwind CSS
- Markdown
- +3 more
Personal Portfolio
This website: a high-performance portfolio built with Next.js 16 and Tailwind CSS 4, with a feature-based architecture, a file-based MDX blog, RSS, structured data, a Redis-backed visitor counter and an automated test and lint pipeline.
- Next.js 16
- React 19
- Tailwind CSS 4
- TypeScript
- Redis
- +1 more
Tools I build with
The core tools behind each build, grouped by what they are used for: cloud and AI, frontend, backend, data, and the platforms they ship on.
Cloud & AI
7 toolsServerless on Lambda, Step Functions, S3, DynamoDB and Textract, as in DocuFlow AI.
Answer generation in the Luxe Wear AI agent platform.
Open models and datasets for experiments.
Agent and retrieval workflows.
LLM integration for agents and automation.
Vector search behind retrieval-augmented answers.
Agent, automation and data experiments.
Frontend
5 toolsCore language for interactive web experiences and modern tooling.
App router, streaming server components, and edge-first mindset.
Component-driven UI architecture and concurrent rendering.
Design system tokens with utility-first ergonomics.
Type-safe patterns, generics, and domain modelling.
Backend
3 toolsMinimal and flexible Node.js web framework for APIs and middleware.
Progressive framework for scalable, enterprise-grade server-side apps.
Lightweight APIs, automation scripts, and platform tooling.
SQL
2 toolsThe world's most popular open-source database.
Relational backbone with JSON support and materialized views.
App
4 toolsContainer workflows, reproducible environments, and CI.
The standard for interface design and prototyping.
Collaboration, reviews, and open-source presence.
Daily driver editor tuned for rapid experimentation.
NoSQL
3 toolsDocument-based database for flexible schema and scalability.
In-memory data store for caching, sessions, and real-time analytics.
Open source Firebase alternative. Postgres, Auth, Realtime.
Platform
2 toolsInstant deployments, managed databases, and effortless scaling.
Edge-first deployment platform with instant previews and analytics.
Open to junior and fresher roles in cloud and AI engineering.
I build serverless systems on AWS and AI agent products, and I like working with teams that care about shipping and verifying what they ship. Also happy to collaborate on open-source or hackathon projects, or just talk agents, cloud and product engineering.