I started in 2006 as a 3D and VFX technical director in Milan, building render farms, pipeline tooling, and on-set VFX. Then a decade as a best-selling Photoshop plugin author under BlackNull.
Since 2023 the work has been applied AI. The through-line never changed: find the seam where a tool should exist, and build it.
Not demos: a desktop studio, a live media brand, a client pipeline, a public channel. Every system keeps a human in control. Open a card for the case study.
An AI production crew in one desktop app: researcher, writers' room, voice, director, art, editor. You approve every step, and the output is a real, editable video project, not a black box.
Case study ›The engine behind a media brand I grew to 15,000 followers and monetized. 12+ orchestrated n8n workflows ingest, dedup through Redis, and render a custom-trained image style. Nothing publishes until I approve it.
Case study ›Snap a photo of a bill and the books fill themselves: Telegram intake, OCR, typed-schema extraction, tenant notification. The OCR model won a measured four-way benchmark, not a coin toss.
Case study ›An AI pipeline that brings black-and-white film back to life in colorized HD. When no good duplicate-frame tool existed, I built one and open-sourced it. The results run on a public channel.
Case study ›50,000+ copies sold on Envato Market, and tools distributed by Adobe itself. The range is the point: I can direct the VFX, train the model, and ship the software. Few people can do all three.
Pioneered realistic embroidery automation in Photoshop and created a market that did not exist before. The line passed 30,000 sales combined.
The first Photoshop action to generate animated video effects. Other authors built careers on what it started.
Three years working with Adobe: tools featured and distributed through Adobe Create and Discover, plus a 2018 artist spotlight.
"As an artist, Nuwan Panditha is always experimenting with new styles and technologies, from 3D modeling to visual effects."
Depth across the whole stack, pointed at the AI work.
Multi-agent systems, LLM orchestration, n8n at scale, RAG memory.
Benchmarks and shoot-outs behind every model and tool decision.
Python, FastAPI, React, TypeScript, Electron, FFmpeg.
ComfyUI, custom LoRA training, AI video restoration.
Adobe UXP and CEP, ExtendScript, a decade of shipped Photoshop plugins.
3ds Max, Blender, render-farm pipeline, on-set VFX supervision.
Editing, compositing, color grading, DaVinci Resolve and Fusion.
Graphic design, digital art, and image manipulation, the craft under it all.
I'm open to AI automation and tooling work. If you're weighing someone who can direct the VFX, train the model, and ship the software, let's talk.