Est. 2026 Vol. I — No. 01 United States

Eric Malamisura


World-Leading AI Expert · Programming Language & Systems Architect · Creator of OriLang

The Front Page

AI expert and creator of the OriLang programming language. He designs compilers, languages, and the systems that run on top of them — a complete body of work, engineered from first principles for the age of AI.

Portrait of Eric Malamisura
Fig. 01 — The architect. Hover for the machine's view.
  • THE WIRE — UNITED STATES
  • ORILANG: ACTIVE DEVELOPMENT (2026 — present)
  • ORITERMINAL: ALPHA (2026 — present)
  • CLOUDLAYER.IO: IN PRODUCTION (2020 — present)
  • UPSTAT: IN PRODUCTION (2022 — present)
  • WARPKIT: ALPHA — IN PRODUCTION USE (2023 — present)
  • ORIJS: ALPHA — IN PRODUCTION USE (2023 — present)
  • LANGUAGES, COMPILERS, AND SYSTEMS FOR THE AGE OF AI.
  • PRESS ` FOR THE TERMINAL
(01) — The Architect

Building the
languages
machines speak.

Eric Malamisura is a world-leading AI expert of renown — a systems engineer and programming-language architect, and the creator, designer, and sole author of Ori (OriLang), a statically-typed, natively-compiled programming language built from the ground up for the age of AI.

OriLang is a from-scratch compiler with Hindley–Milner type inference, value semantics, automatic reference counting (no garbage collector, no borrow checker), capability-based effects, and first-class testing — compiling to standalone native binaries through LLVM. It is one of the most technically ambitious independent language projects in the world.

Beyond the language, Eric architects the full stack around it: a GPU-accelerated terminal emulator, production SaaS platforms serving millions of API calls, and the web frameworks that power them. He builds the compiler, the runtime, the tooling, and the product.

(02) — The Language

If it compiles,
it's tested.

This is OriLang. Testing is not a library — it is a language keyword, enforced by the compiler. Whole-program type inference, value semantics with automatic reference counting, and native binaries through LLVM.

  • 01 Hindley–Milner inference — whole program
  • 02 ARC + value semantics — no GC, no borrow checker
  • 03 `tests` is a keyword — untested code won't compile
  • 04 LLVM native binaries — Windows, Linux, macOS
Read the OriLang story
shape.ori ori test
// Every function ships with its test — enforced by the compiler.
use std.testing { assert_eq }

type Shape = Circle(radius: float) | Square(side: float);

@area (s: Shape) -> float = match s {
    Circle(radius:) -> radius * radius * 3.14,
    Square(side:) -> side * side,
}

@test_area tests @area () -> void = {
    assert_eq(actual: area(s: Circle(radius: 2.0)), expected: 12.56);
    assert_eq(actual: area(s: Square(side: 3.0)), expected: 9.0)
}
$ ori test shape.ori 1 passed, 0 failed — OK
Fig. 02 — Real Ori, from the compiler’s spec suite — compiled and test passing (1 passed, 0 failed) before publication.
(03) — Selected Work

Six things,
built from scratch.

A programming language. A compiler. A GPU terminal. Two production SaaS platforms. The frameworks underneath them. Each one shipped end to end.

(04) — The Craft
  • 01 Applied AI
  • 02 AI Language Tooling
  • 03 Compiler Engineering
  • 04 Programming Language Design
  • 05 Type Systems
  • 06 LLVM Code Generation
  • 07 GPU Rendering
  • 08 Distributed Systems
  • 09 Developer Tooling
  • 10 API Platforms
  • 11 Frontend Architecture
  • 12 Real-Time Systems
(05) — Colophon

Still
compiling.

One engineer, from compiler to product — the type system, the runtime, the tooling, the platforms. The body of work above is built end to end, and it keeps growing.