Mustafa AsghariBrisbane, Australia
Background colour
Mustafa Asghari

AI systems / backend / product engineering

01 Introduction

AI systems-focused software developer building LLM-powered products, source-grounded tutoring systems, high-performance APIs, and production-style backend platforms.

Location
Brisbane, Australia
Focus
AI systems & backend architecture

Building intelligentsystems that think,retrieve, and act.

Explore selected work

02 Selected work

01 Case study

Controlled action

Planning and capability boundaries before execution.

Nexus

Local-firstmacOS AI assistant

Nexus converts voice or text requests into structured action plans and executes them through explicit tools rather than uncontrolled model actions.

  • Deterministic planning
  • Explicit tool execution
  • Privacy-conscious local-first design
  • Memory and orchestration
  • Browser and Apple ecosystem integration
  • Safety controls and capability boundaries

02 Case study

GrindUp

AI-native learning platform

GrindUp is an end-to-end learning platform for subject generation, tutoring, spaced repetition, homework workflows, analytics, and coding practice.

  • Source-grounded tutoring
  • Spaced repetition
  • Sandboxed multi-language code execution
  • Learning analytics
  • Curriculum and quiz generation
  • Structured educational workflows

03 Case study

KabulSweets

Production-stylecommerce platform

Kabul Sweets is a commerce platform with a customer storefront, admin dashboard, ML-assisted pricing interface, backend services, background workers, payments, media processing, and real-time operational updates.

  • Storefront and administration
  • Payments and order management
  • Background processing
  • Real-time operations
  • Media workflows
  • Applied ML support

04 Case study

Music

Private, self-hostedmusic platform

Music extends Koel into a private listening platform, connecting local and online audio with native playback, controlled access, background media pipelines, and a containerized Linux stack.

  • Private access and account approval
  • Native streaming and media caching
  • Background ingestion workers
  • Storage-bounded automation
  • Audio-feature indexing
  • Observability and tested recovery

03 What I build

I build complete systems that turn models, data, and tools into reliable product outcomes.

Learning software development since mid-2024 and building end-to-end projects since late 2024, I focus on clarity, maintainability, performance, safety, and work that is useful beyond a demo.

04 Technical strengths

AI engineering
Agents, tool planning, RAG, embeddings, source grounding, local LLM integration
Backend architecture
FastAPI, modular APIs, async processing, authentication, data modelling
Data & infrastructure
PostgreSQL, Redis, ClickHouse, Typesense, Docker, Alembic, WebSockets, S3
Languages & frameworks
Python, TypeScript, JavaScript, SQL, C#, Java, Next.js, React, Prisma, SQLAlchemy, Celery, OpenAI SDK
Developer tools
Git, GitHub, Vercel, Railway, API design, debugging, system design

05 Education

Queensland University of Technology

Bachelor of Information Technology / Bachelor of Mathematics
Expected 2030

Focusing on artificial intelligence and machine learning, supported by mathematics, statistics, software development, and data analysis.

TAFE Queensland, South Bank

Diploma of Advanced Programming
Graduated 2026

Software development, C#, Java, JavaScript, databases, REST APIs, algorithms, object-oriented design, and practical project delivery.