Identity

Robby Aliasa Akbar

AI Systems Builder & Experimenter

Exploring what AI can do.
Building what works, and are they REALLY useful for everyday life?

Local LLMs, API llms, Cloud LLMs, and Manymore.

I also implement business automation leveraging AI and n8n workflows to streamline customer service and operational administration.

I'm a self-taught AI systems builder and experimenter. I explore local LLMs, inference, automation, agent architecture, and practical AI systems with a focus on finding what actually works.

Positioning

I'm an independent AI Systems Builder and Experimenter.

I'm not here to claim mastery over every AI technology. I'm here to explore what emerging AI capabilities can actually do, test them under real-world conditions, and build systems that genuinely work.

Exploring what AI can do. Building what works. And asking one more question: is it REALLY useful in everyday life?

Journey

AI USER

Starting Point

Starting as a regular AI user without deep technical understanding. Using AI tools to solve everyday tasks, without understanding the architecture behind them.

AI EXPLORER

Deepening

Moving from using AI to understanding how AI systems work. Exploring local LLMs, model architecture, quantization, inference engines, and hardware requirements.

AI BUILDER

Creating

Building integrated systems using models, tools, automation, memory, and infrastructure. Moving from running models to orchestrating complete AI workflows.

AI SYSTEMS EXPERIMENTER

Current

Testing limits, measuring behavior, investigating bottlenecks, and refining systems based on evidence. Focused on understanding what works in real conditions.

Focus Areas

Local AI & Inference

Model deployment, quantization, runtime behavior, CPU/GPU configuration, performance measurement.

AI Systems & Infrastructure

Integration between model, interface, automation, tools, memory, and workflow.

Agent Architecture

Systems that can accept goals, use tools, retrieve information, reason, act, and iterate.

Hardware-Aware AI Evaluation

Understanding AI systems as a full stack: hardware, OS, runtime, engine, model, context, and application.

Hardware-Agnostic Approach

I approach AI systems with a strong focus on adaptability across different hardware environments.

I'm not tied to a specific GPU vendor, device class, or platform. Instead, I focus on understanding how the entire system behaves — from hardware and drivers to inference engines, models, and workloads — and finding ways to make AI work within the constraints of what's actually available.

This includes evaluating:

CPU and GPU behavior
Memory and bandwidth limits
Runtime compatibility
Inference configuration
Model architecture
Context length
Thermal and power behavior
Stability and usability

Principles

Evidence is Evidence, Not a Conclusion

Observations must be understood in context. Experimental results are separated into: observation, interpretation, hypothesis, and conclusion.

Newbie Mindset

Assumptions are treated as hypotheses. When evidence contradicts understanding, the mental model changes.

AI as a Building Partner

AI is used as a partner for exploration, debugging, coding, research, and documentation. But AI output is still tested against documentation, experiments, and actual system behavior.

Coding as a Tool

Coding is used to understand, modify, automate, and control AI systems, not as the final goal.

Value Proposition

AI Feasibility Assessment

Evaluating whether an AI idea is technically realistic and how to approach it.

AI System Architecture

Designing end-to-end AI systems based on real requirements.

Hardware-Aware Optimization

Choosing model, quantization, runtime, and inference configuration based on environment constraints.

AI Workflow & Agent Development

Connecting LLMs with tools, knowledge, memory, automation, and external services.

Technical Experimentation

Testing capabilities, limitations, and trade-offs before committing to larger implementation.

I don't claim to have all the answers in AI.

What I do believe in is my ability to find them by asking the right questions, experimenting, testing assumptions, learning from failures, and validating what actually works. When something works, I turn it into a practical system.

View Experiments