Skip to content
View joyboy257's full-sized avatar

Highlights

  • Pro

Block or report joyboy257

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.

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
joyboy257/README.md

Building AI-native software, automation systems, and data products that solve real operating problems


About me

I’m a Computer Science student specialising in Artificial Intelligence and Machine Learning, focused on building practical systems that connect AI, backend engineering, product thinking, and business operations.

My strongest work sits at the intersection of:

  • AI orchestration and agentic workflows
  • Multi-tenant backend systems and APIs
  • Automation, scraping, and intelligence pipelines
  • Data products and operational dashboards
  • Human-in-the-loop workflows, guardrails, and reliability

I’m most interested in software that moves beyond demos: systems that are useful, testable, observable, and designed around real users.


Current focus

Right now, I’m building and experimenting with AI-native platforms that combine:

  • deterministic backend workflows
  • LLM-assisted orchestration
  • human approval and escalation flows
  • tenant-safe SaaS architecture
  • data pipelines for competitive and operational intelligence
  • product interfaces that non-technical users can actually operate

I care about building systems that are not just clever, but clear, reliable, and production-minded.


Featured work

🚀 Mirai

WhatsApp-native, multi-tenant Revenue OS for SMB operators.

Mirai combines deterministic backend workflows, tenant-safe APIs, operator tooling, and AI-assisted orchestration inside a production-oriented pnpm monorepo.

Highlights

  • service-oriented architecture: API gateway, core backend, workers, realtime inbox, AI orchestrator
  • multi-tenant backend logic with explicit tenancy rules
  • operator dashboard and public marketing site
  • contract tests, security gates, governance checks, and E2E workflows
  • AI assists orchestration while backend data remains the system of record

Stack: TypeScript, NestJS, Next.js, PostgreSQL, Redis, Docker, pnpm, AI orchestration

🏍️ Mototiam

Motorcycle deal finder platform for Singapore.

A full-stack scraping and alerting system that monitors Carousell motorcycle listings, scores deal quality, stores structured data, and alerts users when strong opportunities appear.

Highlights

  • Carousell scraping pipeline
  • dynamic deal scoring and filtering logic
  • PostgreSQL data storage
  • FastAPI backend and REST API
  • Next.js dashboard for listings and top deals
  • Telegram bot for real-time deal alerts

Stack: Python, FastAPI, PostgreSQL, BeautifulSoup, Next.js, Telegram Bot API

Competitive intelligence platform for the Singapore wellness market.

AMI gives a live, evidence-backed view of competitor pricing, offers, SEO footprint, CTAs, and market movements.

Highlights

  • competitor domain discovery from search results
  • website crawling and content extraction
  • pricing, offer, CTA, and SEO signal extraction
  • 0–100 competitor scoring model
  • executive dashboard with market maps, battlecards, offer analysis, and change radar
  • daily automation workflows for freshness and monitoring

Stack: PostgreSQL, n8n, Next.js, Tailwind CSS, SQL, automation workflows

Social media management and scheduling platform.

A collaborative publishing platform for teams managing content across social media channels, with scheduling, platform integrations, analytics, and role-based team workflows.

Highlights

  • authentication with credentials and Google OAuth
  • TikTok OAuth integration and connected accounts
  • post composer with media upload and scheduling
  • publishing infrastructure with audit logs
  • analytics, history, pagination, and error reporting
  • multi-team collaboration with ADMIN, EDITOR, and VIEWER roles

Stack: Next.js 14, TypeScript, Prisma, Tailwind CSS, NextAuth.js

♠️ PokerAI

Reinforcement learning project for No-Limit Texas Hold’em.

A final-year AI project focused on training and evaluating an RL poker agent named Abel through self-play and benchmarking.

Highlights

  • Q-learning based agent updates
  • self-play training framework
  • scripted benchmark opponent, Kane
  • performance metrics tracking
  • evaluation visualizations and logs
  • curriculum-learning oriented experimentation

Stack: Python, Reinforcement Learning, Q-learning, simulation, metrics, visualisation

🧠 AgentOS

Concept-stage AI employee / agent harness.

A product exploration around making persistent, durable AI agents accessible to non-technical business users through natural-language setup, visual monitoring, memory, and escalation flows.

Highlights

  • natural-language-to-agent deployment concept
  • visual agent harness and reasoning traces
  • persistent memory and judgment loop design
  • escalation-first human-in-the-loop model
  • roadmap for Gmail, Calendar, and business data integrations

Stack: AI agents, product systems design, workflow orchestration, agent UX


More projects worth checking

  • Customer Data API — FastAPI REST API for customer data analysis from Excel and database sources.
  • Onyx AI — AI and automation experiments connected to business operations and agency workflows.
  • AMI — market intelligence and automation work related to competitive monitoring.

How I like to build

  • start from the real operating problem
  • prefer clear contracts over magic
  • use AI to accelerate work, but keep humans in the loop where judgment matters
  • design for observability, repeatability, and explainability
  • treat documentation and proof artifacts as part of the product
  • connect technical depth with commercial and product context

Tech stack

Python TypeScript JavaScript PostgreSQL FastAPI NestJS TensorFlow React Next.js Docker Git GitHub

Core: Python, TypeScript, SQL, JavaScript
Backend: FastAPI, NestJS, PostgreSQL, REST APIs, Prisma
AI / ML: Reinforcement Learning, Q-learning, LLM workflows, LangChain / LangGraph familiarity
Frontend: Next.js, React, Tailwind CSS, shadcn/ui
Data / Automation: scraping, n8n, workflow automation, dashboards, Looker Studio
Tooling: Docker, GitHub, CI gates, contract tests, monorepo workflows


What I’m looking for

I’m especially interested in roles where I can work on:

  • AI systems that support real business decisions
  • backend-heavy product engineering
  • data and intelligence pipelines
  • agentic workflows with strong guardrails
  • full-stack SaaS products with meaningful operational complexity

Beyond code

I also spend a lot of time thinking about:

  • product strategy
  • GTM and commercial thinking
  • operating systems for teams
  • how AI should fit into real workflows, not just sit on top as a novelty layer

Popular repositories Loading

  1. optopsy optopsy Public

    Forked from goldspanlabs/optopsy

    A nimble options research and backtesting library for Python

    Python 1

  2. supreme-funicular supreme-funicular Public

  3. cust_data_api cust_data_api Public

    Python

  4. pokerai-public pokerai-public Public

    Python

  5. schedura-app schedura-app Public

    TypeScript

  6. ami-awhl ami-awhl Public

    AWHL Market Intelligence

    TypeScript