Philosophy

Technology changes constantly.What remains is the ability tounderstand, adapt, and create.

System
Human + Machine Learning
Focus
Understanding → Building → Evolving
Status
Always learning
Principle
Delivering a build to perfection
Evolution map

From writing programs to building intelligent systems.

Each node is a stage where the way I approached problems changed, not just the tools I happened to be using.

  1. Where it started

    University Beginning

    Started programming with C++ and learned fundamentals of logic building.

    • C++
    Illustration: a student at a desk piled with C++, Data Structures and Algorithms books, working through confusing code.
  2. The groundwork

    Building Foundations

    Explored Assembly, SQL, data structures, and algorithms, the base underneath everything built since.

    • C++
    • Assembly
    • SQL
    Illustration: a three-monitor desk setup showing C++, Assembly and SQL side by side, with a fundamentals checklist.
  3. Shipping real systems

    Full Stack Development

    Started building complete systems: websites, APIs, databases, automation tools.

    • React
    • FastAPI
    • JavaScript
    • TypeScript
    A code editor showing application source code.
  4. Learning how machines learn

    Entering Artificial Intelligence

    Moved towards machine learning, deep learning, and neural networks.

    • Python
    • TensorFlow
    • PyTorch
    Illustration: working through neural networks and training problems with Python, TensorFlow and PyTorch.
  5. Systems that act

    AI Systems & Agents

    Exploring AI agents, LLM applications, and intelligent automation.

    • LLMs
    • FastAPI
    • RAG
    Illustration: designing an agent workflow with LLMs, RAG and FastAPI alongside an AI assistant.
  6. What comes next

    Future Exploration

    Exploring the future of intelligent machines: the long-term intersection of artificial intelligence, robotics, and autonomous systems.

    Illustration: sketching future possibilities across AI, robotics, autonomous systems and intelligent cities.
Research notebook

How I think, not just what I built.

  1. The starting point

    When I first entered university, programming was completely unfamiliar. I started with C++ without knowing how software was actually built.

    no map, just a compiler

    method.loop
    while (learning) {
      understand();
      build();
      improve();
    }

    Every stage in the map above ran this loop. The tools changed each time. The loop didn’t.

  2. The first principle

    The biggest challenge was not learning syntax. It was learning how to think.

    syntax is the easy half

    learning.py
    def learn(concept):
        while not understood(concept):
            study(concept)
            build_with(concept)
        return teach(concept)
        # teaching exposes the gaps
  3. AI as a learning partner

    As AI evolved, it changed the way I approached technology. Instead of replacing learning, AI became a mentor that helped me explore concepts, debug ideas, and build systems faster.

    a mentor, not a shortcut

    ~/research
    $ python -m venv .venv && source .venv/bin/activate
    $ pip install torch transformers fastapi
      torch ............. [====================] ok
      transformers ...... [====================] ok
      fastapi ........... [====================] ok
    $ jupyter lab --no-browser
      [I] Jupyter Server running
    $  
  4. Teaching as research

    By helping other students solve programming problems, I discovered that teaching others exposed new questions and helped me grow further.

    explaining it reveals the gaps

    teaching.py
    1def teach(student):
    2 explain(concept) # breaks down complex ideas
    3 solve(problems) # discovers new questions
    4 discuss(approaches) # different perspectives
    5 learn_together() # teaching is a two-way process
    6
    7 return growth()
    8
    9# By helping other students solve programming problems,
    10# I discovered that teaching others exposed new questions.
    11
    12# explaining it reveals the gaps
  5. Current belief

    The future of technology is not about memorizing hundreds of tools. It is about understanding concepts, solving problems, creating ideas, and knowing how to use intelligent systems effectively.

    concepts outlive tools

    >_beliefs.sh
    1echo "My current belief:"
    2echo ""
    3echo "The future of technology is not about memorizing tools."
    4echo "It is about understanding concepts, solving problems,"
    5echo "creating ideas, and using intelligent systems well."
    6echo ""
    7
    8# Because the goal is not to use technology,
    9# but to create meaningful solutions with it.
Method

How I build intelligent systems.

One continuous loop, not a checklist. Each phase hands something to the next, and the last one hands it back to the first.

  1. Understand

    Concepts before frameworks.

    • PythonPrimary language for models, data work and backend services.
    • C++Where the fundamentals were learned: memory, types, systems thinking.
    • AlgorithmsComplexity and correctness reasoned about before optimising anything.
    • Data StructuresChoosing the right shape for the data before writing the logic.
    • MathematicsThe linear algebra and statistics sitting underneath every model.

    Every system starts with understanding the problem. Focus on fundamentals, architecture, algorithms and reasoning before selecting tools.

  2. Experiment

    Ideas become prototypes.

    • JupyterNotebook for exploring data and testing an idea before committing to it.
    • PyTorchBuilding and training neural networks.
    • TensorFlowTraining and serving deep learning models.
    • Hugging FacePretrained models and datasets to start from instead of from scratch.
    • OpenAI APIsLanguage-model capability called as a service.

    Build small experiments, test assumptions, and turn concepts into working AI solutions before committing to an architecture.

  3. Engineer

    From prototypes to reliable systems.

    • ReactComponent-driven user interfaces.
    • Next.jsRouting, rendering and the build pipeline for the frontend.
    • TypeScriptTypes that catch mistakes before they ever run.
    • FastAPITyped Python APIs with validation built in.
    • PostgreSQLRelational store for structured application data.
    • DockerThe same environment in development as in production.
    • LinuxThe environment nearly all of this actually runs on.
    • GitVersion history, every change reversible, nothing lost.
    • GitHubWhere code is hosted, reviewed and collaborated on.

    Turn experiments into scalable applications with clean architecture, APIs, databases, containers and a professional development workflow.

  4. Evolve

    Systems improve through feedback.

    • AWSCloud infrastructure to deploy and scale on.
    • CI/CDAutomated checks and deploys on every single change.
    • AI AgentsSystems that plan and act, rather than only respond.

    Continuously learn, monitor, improve and explore emerging technologies: the stack underneath a working system moves faster than the system does.

  5. Teach & Share

    Knowledge grows when shared.

    • ExplainingWalking someone else through it, the fastest way to find your own gaps.

    Teaching programming and explaining concepts reveals deeper understanding and improves engineering thinking. The loop restarts here with a better starting point.

↺ Teach & Share feeds back into Understand. The loop restarts with a better starting point each time.

The system, connected

Nothing here works in isolation.

Drag a node, or hover to trace what it touches.