BUILDER MAP
Builders Around The World
Engineers and researchers whose open code, models and courses shape how AI gets built, with a shelf for Southeast Asia. Pick a name to find them on the globe.
- Countries
- 19
- Cities
- 27
- Builders
- 64
Haleclipse builds open-source tooling around AI coding agents, from the CCometixLine statusline for Claude Code in Rust to the Claudix editor extension and the CometixCode terminal client.
Chip Huyen writes and teaches about putting machine learning into production, from a Stanford systems course to two widely read O'Reilly books.
Sebastian Raschka teaches how language models work by implementing every component in plain PyTorch, backed by the long-running Ahead of AI research newsletter.
Thomas Wolf helped start the Transformers library, which became the default way to share and load pretrained models, and leads science and open research at Hugging Face.
Jerry Liu created LlamaIndex to connect language models with private documents, and the project grew into a document-processing platform for AI agents.
Tri Dao designed FlashAttention, the IO-aware attention kernel now standard in large-model training and inference, and co-created the Mamba state-space architecture.
Tim Dettmers pioneered low-bit quantization methods that let large language models be fine-tuned and served on consumer GPUs.
Arthur Mensch leads Mistral AI, the European lab known for efficient models with openly released weights such as Mistral 7B and Mixtral.
swyx popularized the term AI engineer and runs Latent Space, a podcast and newsletter for developers building on foundation models.
Hamel Husain teaches practical evaluation of LLM applications, contributes to nbdev, and co-authored the CodeSearchNet benchmark.
Jay Alammar turns machine learning papers into visual explainers, most famously The Illustrated Transformer.
Andrew Ng has taught machine learning to millions through online courses and helped found Google Brain, Coursera and DeepLearning.AI.
Julien Chaumond co-founded Hugging Face and oversees the Hub, where the open model community publishes weights, datasets and demos.
Ross Wightman maintains timm, the go-to collection of pretrained PyTorch image models and training scripts.
Aidan Gomez co-authored the original Transformer paper as a Google Brain intern and went on to co-found the enterprise AI company Cohere.
Robin Rombach co-developed latent diffusion, the method behind Stable Diffusion, and later co-founded Black Forest Labs, maker of the FLUX image models.
David Ha explores nature-inspired AI, from the World Models paper to evolutionary model merging at Sakana AI.
Prosper Otemuyiwa is a long-time developer advocate and open-source author, known for Laravel packages such as the Paystack payments integration.
Luciano Ramalho wrote Fluent Python, a respected guide to idiomatic Python, and is a long-time figure in the Brazilian Python community.
Michael Cheng started Engineers.SG, a volunteer effort that records and archives talks from Singapore tech meetups, and helps run several local developer communities.
Khalil Nooh leads Mesolitica, the startup behind the MaLLaM Malaysian language model and voice agents that understand local dialects.
Mior Muhammad Zaki maintains Orchestra Testbench, the standard harness for testing Laravel packages, and contributes to the core Laravel framework.
Charin Polpanumas co-developed WangchanBERTa and other open Thai language resources that much of Thai NLP research builds on.
Vu Anh maintains Underthesea, an open-source Vietnamese NLP toolkit covering word segmentation, tagging, classification and more.
Noval Agung Prayogo wrote Dasar Pemrograman Golang, a free open-source book from which many Indonesian developers learn Go.
Sandhika Galih teaches web development in Indonesian through the WPU channel and courses, reaching a very large audience of beginners.
Jan Christian Blaise Cruz released some of the first open benchmark datasets and pretrained transformer models for Filipino, and co-authored SEACrowd.
nanoGPT open source
A minimal, readable codebase for training and fine-tuning GPT-style models, widely used as a teaching reference.
By
llama.cpp open source
Plain C/C++ engine for running quantized large language models locally on CPUs, GPUs and Apple silicon; home of the GGUF format.
By
Datasette open source
Open-source tool for exploring and publishing SQLite data as interactive websites and JSON APIs.
By
fastai open source
High-level deep learning library on top of PyTorch that powers the free Practical Deep Learning for Coders course.
By
nbdev open source
Literate-programming system for writing, testing, documenting and publishing Python packages from Jupyter notebooks.
By ,
Keras open source
Multi-backend deep learning API for fast, human-friendly model building on JAX, TensorFlow or PyTorch.
By
ARC-AGI open source
The Abstraction and Reasoning Corpus, a benchmark of grid puzzles that tests skill acquisition rather than memorized knowledge.
By
CS 329S: Machine Learning Systems Design course
Stanford course on designing, deploying and maintaining machine learning systems in production.
By
LLMs from Scratch open source
Step-by-step PyTorch code for building, pretraining and fine-tuning a ChatGPT-like model, companion to the book of the same name.
By
Lil'Log course
Long-form technical blog with survey-style posts on agents, diffusion models, reinforcement learning and LLM safety.
By
Transformers open source
Hugging Face library defining thousands of pretrained text, vision and audio model architectures behind one common API.
By ,
LangChain open source
Framework for building LLM-powered applications and agents from composable components.
By
LlamaIndex open source
Framework and platform for connecting language models to private documents through parsing, indexing and retrieval.
By
FlashAttention open source
Fast, memory-efficient exact attention kernels that cut GPU memory traffic during training and inference.
By
bitsandbytes open source
PyTorch library for 8-bit and 4-bit quantization that lets large models fit on smaller GPUs.
By
Mistral AI models product
Family of efficient language models, many released with open weights, including Mistral 7B, Mixtral and Codestral.
By
Next.js open source
React framework for full-stack web apps with server rendering, routing and edge deployment.
By
Terraform product
Infrastructure-as-code tool for declaring and provisioning cloud resources across providers (source-available since 2023).
By
Ghostty open source
Fast, cross-platform terminal emulator with platform-native UI and GPU-accelerated rendering.
By
Photo AI product
Solo-built service that generates AI photos of people from a small set of uploaded selfies.
By
Latent Space product
Podcast and newsletter on the craft of AI engineering, with long interviews of model and tool builders.
By
AI Evals For Engineers & PMs course
Cohort course on error analysis and evaluation design for LLM applications.
By
The Illustrated Transformer course
Visual walkthrough of the Transformer architecture used in machine learning courses around the world.
By
DeepLearning.AI course
Education company offering specializations and short courses on machine learning and generative AI.
By
ImageNet paper
Large labeled image dataset whose annual challenge sparked the deep learning boom in computer vision.
By
PyTorch open source
Open-source deep learning framework with eager execution and GPU acceleration, used across research and industry.
By
timm (PyTorch Image Models) open source
Large collection of pretrained PyTorch image encoders with training, evaluation and export scripts.
By
Ollama open source
Tool for downloading and running open language models locally with a single command and a simple API.
By
vLLM open source
High-throughput LLM inference and serving engine built around PagedAttention memory management.
By
XGBoost open source
Scalable gradient-boosted decision tree library and a long-time favorite for tabular data problems.
By
Qwen open source
Alibaba's family of open-weight language and multimodal models released in many sizes.
By
Attention Is All You Need paper
The 2017 paper that introduced the Transformer architecture underlying modern language models.
By
Stable Diffusion open source
Latent diffusion text-to-image model whose open release brought high-quality image generation to consumer hardware.
By
FLUX open source
Black Forest Labs family of text-to-image models, with open-weight variants for local use.
By
spaCy open source
Industrial-strength natural language processing library for Python with fast pipelines for many languages.
By
OpenClaw open source
Open-source personal AI agent that runs on a user's own machine and acts across chat apps and tools.
By
World Models paper
Paper and interactive article showing agents that learn a compressed model of an environment and train inside imagined rollouts.
By
listmonk open source
Self-hosted, single-binary newsletter and mailing-list manager.
By
Diffusers open source
Hugging Face library of pretrained diffusion pipelines for image, video and audio generation.
By
Elixir open source
Dynamic, functional language on the Erlang VM built for scalable, fault-tolerant systems.
By
Masakhane open source
Grassroots research community building NLP datasets and models for African languages.
By
Laravel Paystack open source
Laravel package for integrating Paystack payments into PHP applications.
By
Unsloth open source
Library and local UI for fine-tuning and running LLMs faster with lower memory use.
By
Fluent Python example code open source
Companion code for Fluent Python, a deep guide to idiomatic Python.
By
SEA-LION open source
Southeast Asian Languages In One Network: open LLMs trained and tuned for the languages of Southeast Asia.
By
Vue open source
Progressive JavaScript framework for building reactive web interfaces.
By
Vite open source
Fast frontend build tool and dev server built on native ES modules.
By ,
Phanpy open source
Minimalistic, opinionated web client for Mastodon and the fediverse.
By
Engineers.SG product
Volunteer-run video archive of talks from Singapore tech meetups and conferences.
By
DragGAN open source
Interactive point-dragging method for precisely editing GAN-generated images (SIGGRAPH 2023).
By
Malaya open source
Natural language toolkit for the Malaysian language, from tokenization to translation and speech.
By
MaLLaM open source
Malaysia Large Language Model: open models pretrained from scratch on Malaysian text.
By ,
Orchestra Testbench open source
Testing helper for building Laravel packages against a full Laravel application.
By
Typhoon open source
Open family of Thai-optimized language, vision and speech models from the SCBX group.
By
PyThaiNLP open source
Python library for Thai natural language processing, from word segmentation to named-entity recognition.
By
WangchanBERTa open source
Pretrained transformer language models for Thai, released with the WangchanBERTa paper.
By
Elysia open source
Ergonomic, end-to-end type-safe TypeScript web framework optimized for Bun.
By
PhoGPT open source
Open generative pretrained language models for Vietnamese.
By
Underthesea open source
Open-source Vietnamese NLP toolkit for word segmentation, tagging, classification and more.
By
TinyUSB open source
Open-source, cross-platform USB host and device stack for microcontrollers.
By
IndoNLU open source
Early large benchmark suite and IndoBERT models for Indonesian natural language understanding.
By
SEACrowd open source
Community data hub and benchmark suite consolidating datasets for Southeast Asian languages.
By , ,
Hacktiv8 product
Indonesian coding bootcamp that trains software developers through intensive programs.
By
Dasar Pemrograman Golang course
Free, open-source Indonesian-language ebook teaching Go from the basics to web development.
By
WPU Course course
Indonesian web development courses from the creator of the Web Programming UNPAS channel.
By
calamanCy open source
spaCy NLP pipelines for Tagalog covering tagging, parsing and named-entity recognition.
By
Filipino Text Benchmarks open source
Open benchmark datasets and pretrained transformer models for the Filipino language.
By
CCometixLine open source
A fast Rust statusline for Claude Code showing the model, folder, git branch and context window, with a TUI to theme it.
By
Claudix open source
A VS Code extension, built on the Claude Agent SDK, that brings a Claude Code chat panel with session history into the editor.
By
CometixCode open source
An unofficial Rust reimplementation of the Claude Code terminal interface, built on the iocraft TUI library.
By
Pins mark only the city each person states on their own public profile; people who state just a country get a hollow marker, and the rest appear in the list without a pin. Avatars are their public GitHub pictures. To suggest a builder or ask for a change, email hello@exepert.com.