From Hollywood VFX to Generative AI
A curated selection of projects showcasing my journey through traditional VFX, animation, and the emerging world of generative AI. These projects represent key milestones in my ongoing exploration of technology and creativity.
TCL Research America
Art Director / Research Supervisor · October 2024 - Present
Leading creative and technical innovation at the intersection of artificial intelligence and visual entertainment. Directing creative development and visual direction for AI-generated video projects, including art direction for "The Audition" as part of TCL's inaugural FILM MACHINE program.
- Leading R&D teams exploring cutting-edge AI tools for animation production
- Developing comprehensive pipeline strategies for generative AI integration
- Driving interdisciplinary collaboration between artistic and technical teams
- Supervising implementation and testing of AI models for animation
- Pioneering best practices for AI-assisted animation workflows
Workflow Integration
Seamless integration of AI models into traditional animation pipelines
Quality Control
Developing systems to ensure consistent output quality
Dune: Part 2
CG Supervisor at DNEG · 2021-2023
Supervised all aspects of the CGI pipeline from matchmove and layout to final lighting and render for several sequences in this Oscar-winning production. Working remotely with teams across India, Vancouver, and Montreal.
- Led CG supervision for the India Team, where the bulk of the VFX work was done
- Oversaw several major sequences including the arena fight, spice miner ambush, and Raban ornithopter attack
- Coordinated complex visual effects across international teams
- Contributed to the film's Academy Award for Best Visual Effects
Global Coordination
Managed teams across multiple international studios
Award-Winning
Contributed to 2024 Academy Award for Best Visual Effects
IntelliStory
Structured Generative AI for Film, Animation, and Art Direction · Beta
IntelliStory brings a full production pipeline to generative AI. Power your creative process from first idea to final render — storyboards, concept art, character design, previs, and final shots, all within one cohesive system.
Concept to Final
Complete Pipeline
Studio Grade
Version Control
Multi-Model
AI Integration
Complete Generative AI Production Pipeline
- Real Production Pipeline: Automatic hierarchy, naming, versioning, and cloud sync. Projects → Sequences → Shots → Assets — built like a VFX or animation studio
- Storyboarding & Previs: Build full sequences and timelines with AI-generated imagery. Visualize your story before production begins
- Character & Asset Design: Manage character designs, variants, and creative iterations seamlessly. Maintain consistency across your entire production
- Generative AI Integration: Run top AI models in parallel for cinematic-quality concept, art, or shot generation. Compare results and iterate instantly
- Team Collaboration: Every prompt, ref, and generation stays organized and shareable. No more messy folders or untraceable prompts
- Transparent Spending: Built-in budget visibility for every project. View generation costs as they occur, monitor team-wide spending
A web-based 3D viewer built on Apple Research's SHARP model. Generates over 1.1 million Gaussian splats from a single image in under 1 second — enabling instant 3D reconstruction with zero optimization.
Zero-Shot
Single Image to 3D
1.1M Points
High-Density Splats
Interactive
WebGL Viewer
Instant 3D Reconstruction
- Apple Research Model: Built on the SHARP monocular view synthesis architecture with DINOv2 backbone
- Real-Time Rendering: Custom point cloud shader with interactive rotation, zoom, and orbit controls
- Sub-Second Inference: Generate complete 3D scenes from any photograph in under 1 second on GPU
- Production-Ready Export: Outputs standard PLY files compatible with Gaussian Splatting renderers
Most "Markdown → PDF" exports still look like raw chat output. I Like Your Docs! fixes that — a premium, 100% client-side converter that produces print-ready PDFs with real typography and consistent icon rendering.
Privacy First
100% Client-Side Processing
Print Ready
Professional Typography
Smart Icons
Emoji to SVG Conversion
Better Than Chat Exports
- Professional Themes: 14+ typography themes including technical, legal, academic, and resume styles
- Smart Icon System: Automatically converts emojis to crisp SVG icons (Lucide, Heroicons, etc.) for professional use
- Visual Control: Split editor with live preview and visible page guides to perfect your page breaks
- Private by Design: Totally free and private. Everything runs locally in your browser with no uploads
HiveMind is a sophisticated multi-LLM orchestration platform that enables seamless coordination between multiple AI models. It provides a unified interface for managing complex AI workflows across different language models.
Key Features
- Multi-Model Support: Orchestrate multiple LLMs simultaneously for enhanced capabilities
- Unified Interface: Single platform to manage diverse AI models and workflows
- Intelligent Routing: Automatically route tasks to the most appropriate model
- Workflow Automation: Create complex AI pipelines with ease
Multi-LLM
Coordinate multiple AI models in one platform
Orchestration
Intelligent task routing and workflow management
MetaProbe is an AI-powered media metadata extraction tool that automatically analyzes and extracts comprehensive metadata from media files. Essential for content management and AI-assisted production workflows.
Capabilities
- AI-Powered Analysis: Intelligent metadata extraction using advanced AI models
- Comprehensive Data: Extract technical and creative metadata from media files
- Workflow Integration: Seamlessly integrate with production pipelines
- Automated Processing: Batch process large media libraries efficiently
AI-Powered
Intelligent metadata extraction
Automated
Batch processing capabilities
MOTH (Machine-Optimized Text Hierarchy) is a compact notation format for technical specifications, PRDs, and coding rules. Designed specifically for AI-assisted development, MOTH achieves 70-90% compression while maintaining full readability.
Why MOTH?
- 70-90% Compression: Dramatically reduce token consumption vs traditional markdown
- LLM-Native: Optimized for AI coding agents and language models
- Human Readable: Maintains clarity despite extreme compression
- Structured Format: Perfect for PRDs, architecture docs, and coding rules
70-90%
Compression vs markdown
LLM Native
Built for AI workflows
LLM Compression Comparison Tool
Interactive tool comparing compression formats optimized for LLM context windows. Compare MOTH, TOON, and kablUI side-by-side with real GPT tokenization (o200k_base).
MOTH
70-90% compression
TOON
40-60% compression
kablUI
60-85% compression
kablUI
Compact UI Notation System · Released 2024
UI Design With Explosive Efficiency
kablUI uses only 5.5% of the characters needed for equivalent HTML, dramatically reducing token consumption in LLM contexts.
Why kablUI Changes Everything
-
95% More Efficient
kablUI uses only 5.5% of the characters needed for equivalent HTML, dramatically reducing token consumption in LLM contexts.
-
Built for LLMs
Optimized for AI processing, making it ideal for design collaboration with language models like GPT-4 and Claude.
-
Human Readable
Intuitive syntax that balances extreme brevity with natural readability for human designers and developers.
-
Enforces Structure
Naturally guides designers toward well-organized, hierarchical interfaces with clear component relationships.
-
Rapid Iteration
Make structural UI changes in seconds instead of minutes, accelerating the design exploration process.
-
Framework Agnostic
Works with any UI framework or design system—from React and Vue to Flutter and native platforms.
@APP[bg:gray-900]
@SB[w:240 bg:gray-800]
§PROJECTS
+Project01
-Project02
-Project03
§SEQUENCES
+seq01
-seq02
§AI TOOLS
-⚡Prompt
-🎨Style
-âś“Consistency
@MAIN
@NAV[h:30 bg:gray-800]
Project01 | seq01
@TABS [Shots][Characters][References]
@STATS "3 Seq | 30 Shots | 12 Done"
@TIMELINE
@CARD[w:40] "seq01_sh001"[blue] "COMPLETE"[green]
@CARD[w:40] "seq01_sh002"[blue] "IN PROGRESS"[yellow]
@PREVIEW
@VIDEO[aspect:16/9]
@CONTROLS ⏮️ ▶️ âŹď¸Ź 🔊
@COMPONENT[Dialog]
@HEADER "Confirm Action"
@BODY
"Are you sure you want to proceed with this action?"
@FOOTER
[Button secondary]Cancel[/Button]
[Button primary]Confirm[/Button]
@COMPONENT
A cyberpunk-themed web application that generates infinite ambient soundscapes inspired by Vangelis' iconic "Blade Runner Blues" from the Blade Runner soundtrack. Features Moog-style synthesis with advanced effects.
Generative
Infinite Ambient Music
Moog-Style
4-Pole Ladder Filter
10 Presets
Curated Moods
Web Audio Synthesis Engine
- Multi-layered Oscillators: 3 oscillators per voice (2 sawtooth + sub) for thick, rich sound
- Moog Model D Emulation: Authentic 4-pole ladder filter for classic analog warmth
- Effects Chain: Master filter → Delay → Reverb with real-time control
- Data Integration: Connect real-time data sources (weather, stocks, time) to control synth parameters
ComfyUI LLM Bridge
Give LLMs Accurate Knowledge of Your Local ComfyUI · Open Source
Compresses ComfyUI's massive /object_info JSON (10MB+) into compact, LLM-optimized notation. Stops hallucinated node names, prevents type misconnections, and saves tokens.
80%+
Compression
Live
Node Discovery
Domain
Splitting
How It Works
- Live Discovery: Fetches node definitions directly from your running ComfyUI instance
- 80%+ Compression: Converts verbose JSON to a single-line notation (@NodeName +input:TYPE -output:TYPE)
- Domain Splitting: Organizes nodes by task (video, 3D, segmentation, API, training)
- IDE Integration: Ready-to-use Antigravity workflow — just run /comfyui
Browser-based tools for converting verbose ComfyUI JSON into compact formats for LLMs. Save 50-80% on token costs with bidirectional compression. All processing happens locally — no data sent to servers.
50-80%
Token Savings
Offline
Local Processing
Visual
Graph Preview
Features
- Node Compressor: Convert verbose node definitions (object_info.json) into compact LLM-friendly format
- Workflow Compressor: FlowZip format uses 50-70% fewer tokens while preserving all workflow data
- Workflow Inflater: Convert FlowZip back to standard JSON for use in ComfyUI
- Graph Visualizer: Interactive workflow preview with pan, zoom, and connection styling
A TypeScript library that transforms casual Unicode emojis into high-quality SVG icons from industry-standard libraries. Over 7,800+ mappings across 10+ icon sets for professional documentation.
7,800+
Icon Mappings
10+ Sets
Icon Libraries
TypeScript
Type-Safe API
Supported Icon Libraries
- Lucide Icons: Modern, clean design for web applications
- Heroicons: Tailwind CSS optimized icon set
- Material Icons: Google's Material Design family
- And More: Font Awesome, Feather, Phosphor, Tabler, Boxicons, Remix
Promptobot is a collection of advanced AI solutions and creative tools designed to enhance productivity and creativity in AI-assisted workflows. It provides powerful utilities for prompt engineering and AI interaction.
Features
- Prompt Engineering: Advanced tools for crafting effective AI prompts
- Creative Solutions: Innovative approaches to AI-assisted creativity
- Workflow Enhancement: Streamline AI interactions and boost productivity
- Versatile Tools: Comprehensive suite for various AI applications
AI Tools
Advanced prompt engineering solutions
Creative
Enhance AI-assisted workflows
