NodeDB
A Rust database project that brings document, vector, graph, columnar, time-series, key-value, full-text, and array workloads into one engine for AI and agent applications.
Open NodeDBAbout Farhan Syah
I build open-source systems for data, machine learning, and design tooling.
I am a Malaysia-based programmer with more than a decade of software work behind me. My current focus is practical infrastructure: databases that can handle modern AI workloads, Rust-native machine learning building blocks, and design tools that agents and humans can edit without fragile binary files.
What I Am Working Toward
I want to spend more of my time on open source: maintaining existing projects, building new tools, and helping other builders learn through practical examples. Sponsorship and community support make it easier to go deeper instead of treating this work as something squeezed between everything else.
The common thread is reducing tool sprawl: fewer moving parts, clearer source files, and systems that are useful to solo builders as well as serious engineering teams.
My Work
Machine Learning in Rust
An ecosystem of Rust crates for machine learning infrastructure, including tokenizer, numerical computing, notebooks, regex/tokenization tools, and ML primitives.
Open ml-rustZenith
A plain-text design document format and renderer for posters, decks, social graphics, books, and other visual documents that should be readable, diffable, and reproducible.
Open ZenithFeatured Open-Source Pieces
splintr
High-performance tokenizer work for BPE and SentencePiece with Rust and Python bindings.
GitHubnumr
Numerical computing experiments for Rust, including GPU-oriented direction inspired by NumPy-style workflows.
GitHubvenus
A Rust notebook direction that keeps work in standard .rs files while supporting fast feedback loops.
GitHub