Skip to content
View TongDucThanhNam's full-sized avatar
  • Ho Chi Minh City

Highlights

  • Pro

Block or report TongDucThanhNam

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
TongDucThanhNam/README.md

Hi, I'm Tong Duc Thanh Nam

AI Product Engineer building agentic workspaces, browser copilots, and AI-native productivity tools.

Email · Eragear · Portfolio · LinkedIn


About Me

I build AI-native software products that combine full-stack engineering, agentic workflows, and practical user experience.

My current focus is Eragear — an AI workspace ecosystem for long-running conversations, artifacts, browser-assisted research, local knowledge workflows, and agent integrations.

I care about:

  • Designing AI products that are actually useful in daily workflows
  • Building streaming chat systems, model routing, credit usage, and artifact experiences
  • Creating agentic interfaces across web apps, browser extensions, local tools, and knowledge systems
  • Shipping full-stack TypeScript products with Cloudflare, Hono, TanStack, Drizzle, Better Auth, and Bun

Main Project: Eragear

Eragear is an AI-native workspace for long-running conversations, in-context artifacts, research workflows, and agent-powered productivity.

Area What I Built
AI Workspace Agent-first chat experience with selectable models, tools, search, and thinking controls
Streaming AI Backend Hono-based AI API with model routing, streaming responses, usage tracking, and persistence
Artifacts Sidecar workspace for AI-generated documents and rich content
Credits & Billing Usage-based credit accounting and payment webhook architecture
Browser Copilot Extension that bridges browsing context into the AI workspace
Knowledge Copilot Obsidian assistant using deterministic graph-based context instead of only vector search
Code Copilot ACP-based coding assistant connecting browser UI with local AI agents

Featured Eragear Projects

Project Description Stack
Eragear Core AI-native workspace with chat, artifacts, credits, billing, auth, and Cloudflare deployment TanStack Start, React, Hono, Cloudflare Workers, D1, Drizzle, Better Auth, Bun
Eragear Web Extension Browser copilot for human-in-the-loop research and verified web context collection WXT, React, TypeScript, Tailwind CSS
Eragear Code Copilot Web-based AI coding assistant using Agent Client Protocol to connect UI with local coding agents React, Hono, tRPC, ACP, Zustand
Eragear Obsidian Copilot Local-first Obsidian assistant using graph-based context retrieval Obsidian API, TypeScript, Graphology, MiniSearch, Web Workers
AI Chatbot Tree Branching conversation UI with conversation-tree SSOT and projection-based context control Next.js, AI SDK, Zustand, D3, React Flow

Technical Focus

AI Product Engineering

  • Streaming AI chat interfaces
  • Multi-model routing and provider abstraction
  • Tool-enabled agents and selected tool control
  • Long-context conversation handling
  • Artifact generation and sidecar workspaces
  • Usage tracking, credits, billing, and rate limits

Full-stack TypeScript

  • React / TanStack Start / Next.js
  • Hono / Cloudflare Workers
  • Drizzle ORM / Cloudflare D1 / SQLite
  • Better Auth
  • Bun / Turborepo / Biome

Agentic Interfaces

  • Browser extension copilots
  • Obsidian/local-first knowledge tools
  • ACP/MCP integrations
  • AI coding assistants
  • Human-in-the-loop research flows

Selected Engineering Highlights

  • Built a Cloudflare-native AI workspace architecture with separate web, server, auth, database, and AI packages.
  • Implemented streaming chat flows with usage metadata, message persistence, and credit deduction.
  • Designed an agent selection system with persisted model, tool, and thinking-budget state.
  • Built browser-extension workflows that let users verify search results before sending context to AI.
  • Explored deterministic knowledge retrieval using graph algorithms such as PageRank and spreading activation.
  • Built conversation-tree UX experiments for branching, regenerating, and composing AI context.

Current Direction

I am currently focused on:

  • Turning Eragear into a polished AI workspace product
  • Improving agentic UX for research, coding, and knowledge work
  • Building reliable AI infrastructure around streaming, billing, persistence, and tool use
  • Creating public case studies from private product work

Contact

I am open to roles related to:

  • AI Product Engineer
  • Full-stack AI Engineer
  • Agentic Workflow Engineer
  • Frontend/Full-stack Engineer for AI products

Email: tongducthanhnam@gmail.com

Pinned Loading

  1. SellerKeyboard SellerKeyboard Public

    Seller keyboard for create Snippets text.

    Kotlin

  2. ai-chatbot-tree ai-chatbot-tree Public

    Tree-based chat UI

    TypeScript 4 1

  3. ai-chat-app ai-chat-app Public

    Chat AI is an AI LLM Services like ChatGPT. It focuses on improving both AI model's response and helping users improve their knowledge:

    TypeScript 1

  4. eragear-code-copilot eragear-code-copilot Public

    Eragear Code Copilot is an Electron-based local AI development environment built around **Supervisos**—an intelligent engineering supervisor that goes beyond ordinary code completion.

    TypeScript

  5. eragear-obsidian-copilot eragear-obsidian-copilot Public

    TypeScript