Main Concourse Open For Boarding

I make AI systems easier to ship, operate, and understand.

I work on model-serving infrastructure, developer tooling, and technical writing that helps teams move from prototype to production.

Focus

Production AI

Strength

Clear Interfaces

Approach

Hands-On

Departure Board

Current Work

PLT 01 Production inference systems Serving
PLT 02 Developer tooling and explainers Shipping
PLT 03 Technical writing for reusable systems work Writing

Systems

Triton, vLLM, Ollama, TensorRT, and LLM product workflows.

Output

Projects, explainers, and notes built to be inspected and reused.

See Projects Read Writing View GitHub
Production Inference Developer Tooling Frontend Explainers
Platform A Operating Notes

Work across inference, interfaces, and technical communication.

The focus is practical AI engineering with enough product and frontend discipline to make the systems useful, observable, and understandable.

Line 01

Inference Systems

Serving, tuning, and simplifying model workflows so product teams can run AI features without guesswork.

Line 02

Interactive Interfaces

Building frontends and explainers that make technical systems easier to inspect and use.

Line 03

Technical Writing

Turning experiments into notes, tutorials, and implementation guides that other engineers can reuse.

Platform B Latest Dispatches

Recent writing on machine learning, engineering, and experiments in progress.

See All Posts
Stop 01

Galton-Watson Process

In this post, we outline the Galton-Watson Process.

Read Dispatch
Stop 02

Policy Iteration Algorithm in a Nutshell

Policy Iteration with Maze Example.

Read Dispatch
Stop 03

Iterative Policy Evaluation in a Nutshell

Check if an agent is doing the right thing with Policy Evaluation Methods

Read Dispatch
Platform C Projects In Motion

Selected projects across machine learning, visualization, tooling, and product experiments.

    Platform 01
    Project
    Race Car RL Agent
    PythonReinforcement LearningOpenAI Gym

    Developed a Deep Q-Learning agent in a simulated racing environment to autonomously navigate tracks by learning from reward feedback.

    Read the Paper
    Platform 02
    Project
    Garbage Classification with CNN
    PythonComputer VisionEfficientNet

    Transfer-learned an EfficientNet backbone to classify multiple waste categories in images, enhancing model performance through data augmentation.

    Read the Paper
    Platform 03
    Project
    Haiku Generator
    LLMsPrompt EngineeringNLP

    Explored prompt-engineering strategies with large language models to generate structured haikus, enforcing syllable constraints and thematic coherence.

    View AppSource Code
    Platform 04
    Project
    D3.js Reference Tutorial
    JavaScriptD3.jsData Visualization

    Interactive web guide demonstrating core D3.js features and visualization techniques.

    Source CodeLive Demo
    Platform 05
    Project
    Chat App
    ElmNode.jsExpress.js

    Real-time chat app with WebSocket messaging; Elm frontend and Express.js backend.

    Source Code
    Platform 06
    Project
    Twitch Python Discord Bot
    PythonDiscord.pyTwitch API

    Python Discord bot with Twitch API integration for live stream alerts and moderation.

    Source Code
    Platform 07
    Project
    Rust Tauri App
    RustTauriDesktop App

    Cross-platform desktop color picker built with Rust and Tauri, showcasing native UI integration.

    Source Code
    Platform 08
    Project
    Guided Sudoku Solver
    AlgorithmsBacktracking

    Interactive Sudoku solver with step-by-step guidance built in Python using a backtracking algorithm and heuristic improvements.

    DemoSource Code
    Platform 09
    Project
    Wordle Solver
    TypescriptEntropyTanStack Start

    Interactive wordle solver using entropy with a practice mode for improvements.

    Demo
    Platform 10
    Project
    Interval Timer
    TypescriptProductivityTanStack Start

    Interval Pomodoro timer for Work/Study Sessions.

    Demo
Platform D Learning Route
Active Route

The roadmap tracks topics worth studying deeply enough to turn into usable systems, notes, or experiments.

Current learning route across reinforcement learning, performance engineering, graph systems, and agents.

  1. Stop 01
    In Progress

    Applied Reinforcement Learning

    Reading

    Reinforcement Learning (Sutton & Barto), Deep Learning (Goodfellow et al.)

    Building With

    MarkdownPythonObsidianPyTorchNumPy

    Exploring

    RL algorithms • Policy gradients

  2. Stop 02
    In Progress

    High-Performance Systems

    Building With

    CUDAPolarsTritonParallelizationTensorRT

    Exploring

    Distributed systems • GPU acceleration

    Operating Principles

    • Prototype fast in Python, then optimize the critical path deliberately.
    • Use hardware acceleration where latency matters enough to justify complexity.
    • Keep operational visibility alongside performance work.
  3. Stop 03
    In Progress

    Graph Intelligence

    Reading

    Graph Representation Learning (Hamilton)

    Building With

    PyTorch GeometricNeo4jNetworkXC++

    Exploring

    Knowledge graphs • GNNs

  4. Stop 04
    In Progress

    Autonomous Agents

    Reading

    Generative Agents (Park et al.)

    Building With

    LangChainOpenAI APIAutoGPTFastAPI

    Exploring

    Multi-agent systems • Emergent behavior

Signal Box Inspiration Sources

Tools and products that influence how I design interfaces and explain systems.

Yard Notes

These are references for clarity, speed, and product restraint rather than trend chasing.

Ref 01

v0.dev

Generative UI prototyping for fast idea validation.

Open Reference
Ref 02

tldraw

Sketch-first thinking for product and systems work.

Open Reference
Ref 03

Linear

Clear execution and issue tracking without clutter.

Open Reference
Ref 04

Excalidraw

Diagramming that stays rough, legible, and quick to share.

Open Reference
Ref 05

NotebookLM

Research synthesis and study support from messy source material.

Open Reference