AI / LLM Engineer · Backend Engineer · Open Source Builder
Building practical systems around LLMs, RAG, AI Agents, Backend Engineering and Developer Infrastructure.
I'm Mustakim, a software engineer focused on building AI-powered applications and backend systems.
My work started primarily around backend development with Java and Spring Boot and gradually moved toward LLMs, RAG, AI agents, NLP and developer infrastructure.
I enjoy working on problems where software engineering and AI overlap:
- Designing backend systems that can scale
- Building applications around LLMs
- Experimenting with RAG and agent architectures
- Improving AI application reliability and evaluation
- Building developer-focused tools and infrastructure
- Turning research ideas into working software
Most repositories here are experiments, prototypes or products that started with a simple question:
Can this be built better?
|
|
Serverless infrastructure and developer tooling.
Exploring a simpler developer experience for building and running serverless applications.
The project focuses on the layer between application code and infrastructure:
Application
|
v
Nova Serve
|
+-------- Functions
|
+-------- Services
|
+-------- Infrastructure
Technology
TypeScript · Node.js · Serverless
LLM dataset engineering and preparation.
A project exploring the workflow required to turn raw data into datasets suitable for LLM fine-tuning and evaluation.
Raw Data
↓
Cleaning
↓
Validation
↓
Transformation
↓
Quality Checks
↓
Training Dataset
Technology
Python · NLP · LLMs · Dataset Engineering
AI-assisted enterprise integration platform.
Exploring how intelligent processing can be introduced into enterprise integrations, APIs and automated workflows.
The main focus is combining:
Enterprise Systems → Integration → AI Processing → Automation
Technology
Python · AI · LLMs · APIs
Infrastructure for experimenting with autonomous AI systems.
Instead of relying on a single general-purpose agent, AgentOS explores specialized agents working together.
Controller
|
+-------------+-------------+
| | |
v v v
Research Planning Execution
| | |
+-------------+-------------+
|
Result
Technology
Python · LLMs · Agents · Orchestration
Multi-agent AI system for autonomous marketing workflows.
AgencyOS explores how specialized agents can collaborate across different stages of a workflow instead of treating an AI application as a single prompt-response system.
Strategy
↓
Research
↓
Content
↓
Review
↓
Approval
↓
Execution
Technology
Python · LLMs · Multi-Agent Systems
AI-powered document learning platform.
PageTutor AI explores how document intelligence and RAG can turn large PDFs into structured learning experiences.
PDF
↓
Document Processing
↓
Structure & Topics
↓
Retrieval
↓
Context Construction
↓
LLM
↓
Learning Experience
Technology
Python · FastAPI · RAG · LLMs · Next.js
AI-assisted root cause investigation.
Exploring how AI can help engineers move from system symptoms to evidence, possible causes and actionable conclusions.
Incident
↓
Evidence
↓
Signals
↓
Possible Causes
↓
Correlation
↓
Root Cause
Technology
Python · AI · LLMs · Root Cause Analysis
LLM Engineering
|
+-- Fine-tuning
+-- Evaluation
+-- Inference
+-- RAG
+-- Memory
|
v
Agentic Systems
|
+-- Planning
+-- Tool Use
+-- Multi-Agent Systems
+-- Autonomous Workflows
|
v
AI Infrastructure
|
+-- Model Serving
+-- Developer Tooling
+-- Observability
+-- Serverless
+-- Performance
I also spend time strengthening my fundamentals through Data Structures & Algorithms and system design.
I also mentor developers and students, particularly around programming, backend engineering and AI.
Teaching has become an important part of how I learn:
Build → Explain → Question → Improve → Build again
If you're working on something around AI, LLMs, backend systems, developer tools or open source, feel free to reach out.
Build systems. Learn deeply. Share openly.