Work with me

Engine performance, rendering, and procedural 3D in one workflow.

I take on focused work for studios and technical teams where performance optimization, graphics engineering, engine architecture, tools, and technical art overlap. The work stays close to the code and production workflow, with scope and success criteria agreed after I understand the project.

The operating advantage

Connect engine and art-pipeline decisions.

Rendering engineers, tools programmers, and technical artists often solve adjacent parts of the same production problem through different assumptions, data formats, and priorities. That handoff becomes expensive when assets, procedural systems, and runtime constraints must evolve together.

I work across that boundary directly: C and assembly systems programming, graphics APIs, engine tooling, Houdini procedural workflows, animation, and game-ready 3D assets. I can trace a bottleneck from source geometry to exported data, runtime memory, rendering cost, and final frame output.

Ways I can help

Three areas of work.

Each engagement begins with the current constraint, a documented baseline, an agreed scope, and a testable outcome.

Pipeline architecture

Procedural 3D & Houdini Pipeline Automation

Replace repetitive asset and level work with explicit procedural systems connected directly to the engine.

  • Procedural geometry and variation systems
  • Automated UV and asset preparation
  • CSV or custom binary level exporters
  • Houdini-to-engine data contracts
  • Validation and batch processing tools
  • Runtime integration and profiling

Focus: identify repetitive stages, automate suitable steps, and improve the path from authoring to runtime.

Performance optimization

Engine & Graphics Performance Optimization

Find and fix the systems responsible for frame-time spikes, CPU or GPU bottlenecks, memory pressure, rendering stalls, slow loading, or expensive iteration.

  • Frame-time capture and hot-path profiling
  • CPU/GPU bottleneck investigation
  • Memory lifetime, allocation, and peak-use analysis
  • Data layout and CPU-cache behavior
  • Render submission and API synchronization
  • Entity, animation, and skeletal-deformation costs
  • Loading, resource, and engine-tool performance
  • DirectX 11 rendering diagnostics

Focus: establish a repeatable baseline, isolate the expensive path, implement the change, and compare the result under the same workload.

Custom software

Custom 3D Software & Simulation Architecture

Build purpose-specific real-time 3D software when a commercial engine introduces more constraints than it removes.

  • Interactive 3D and simulation applications
  • Domain-specific editors and debuggers
  • Visualization and digital-twin systems
  • Custom import, export, and processing tools
  • Rendering and interaction architecture
  • Long-term technical ownership plans

Focus: build a focused native tool around agreed startup, integration, delivery, and source-ownership requirements.

Watch the systems being built

Architecture you can inspect before the first call.

Start with the project overview, then continue into the detailed architecture and pipeline breakdowns.

Project overview

Sange: gameplay, engine, and development workflow.

An overview of the project covering gameplay, Njinn’s architecture, the rin language, and the art workflow.

Watch on YouTube →

Engine architecture

Handmade Game Engine Architecture

Deep dive into Njinn’s core loop, data-oriented design, C arena allocators, rendering, state machines, physics, generics, and assembly optimization.

Watch on YouTube →

Procedural production

Automated 3D Art & Procedural Level Pipeline

A closer look at the procedural asset and level-generation workflow used while developing Sange.

Watch on YouTube →

Selected projects

Systems you can inspect in a working project.

Sange

Public demo on a custom stack

Sange is an action roguelike in development with a public Steam demo. It uses Njinn for procedural rooms, combat state machines, skeletal and boss animation, rendering, content tools, and the Houdini asset pipeline.

What it demonstrates: engine and content-pipeline decisions can be inspected together in a playable project.

Njinn

Explicit memory and runtime performance

Njinn uses permanent, transient, and level-scoped arena allocation; data-oriented entity systems; skeletal deformation; immediate-mode tools; and a focused rendering pipeline.

What it demonstrates: allocation strategy, data layout, entity updates, and rendering costs remain visible for profiling and optimization.

rin

Language tooling built around the engine workflow

rin transpiles to readable C and provides monomorphized generics, compile-time reflection, source-level debugging, and arena-aware standard containers.

What it demonstrates: the compiler, generated C, and source-level debugging can be inspected together.

Engagement structure

How we work together.

Choose the operating model that matches the scope, urgency, and ownership your team needs.

Project intake

Bring me the bottleneck.

Send the current constraint, the system it affects, and the outcome your team needs. I’ll reply with whether the problem fits a focused architecture sprint, a monthly technical retainer, or a custom build.

Opens a structured email draft. No form service, account, or scheduling gate.