Hi, I'm

Nguyen Phong Pham

Product Manager • AI & Cloud

I build digital products and explore technology to create meaningful experiences and solve real-world problems.

Contact for work →
Nguyen Phong Pham - Watercolor Portrait

About

I'm a product manager with a passion for technology, user experience and building solutions that make a positive impact.

Experience

Ecomdy
Product Manager Ecomdy
2022 – 2026 TikTok Agency
Green Global
Project Manager Green Global
2017 – 2022 IT Solutions
Cenvisco
Project Manager Cenvisco
2013 – 2017 IT Consulting
Gameloft DAD
Game Designer Gameloft DAD
2011 – 2013 Game Studio

Contact

Feel free to reach out if you'd like to collaborate or just say hello.

•
LOCAL AI COMMAND CENTER

Phong's AI Studio & Multi-Project Hub

Unified dashboard for managing multiple projects and directly commanding your local RTX 5070 AI models.

P Mavinhthanh@gmail.com
RTX 5070 (12GB VRAM)
ComfyUI :8188
Qwen LLM :8081

Active Projects Portfolio

ACTIVE PRODUCTION
🎬

Mun: Guardian of the World Tree

Anime Dark Fantasy • 8 Episodes

Comprehensive generative anime film production. Dark Fantasy Impasto Oil Painting style, powered by MiniMax-H3, Wan 2.1, and Gemini TTS.

Current: Episode 8 Master Render 80%
MiniMax-H3 • Illustrious XL
📢

Halwa Ad Campaigns & Video Ads

Commercial & Motion Graphics

Automated script generation, Google Cloud TTS voiceover, and video assembly for commercial advertising campaigns.

Status: 3 Commercial Variants Built 100%
GCloud • FFmpeg
🎨

Photorealism & Creative Art Lab

Digital Art & Portrait Generation

High-fidelity uncensored photorealism, anatomy exploration, and studio lighting utilizing CyberRealistic XL, PonyRealism, and FLUX.2 + Lustly LoRA.

Model Presets: 4 Checkpoints Active Ready
CyberRealistic • PonyRealism
⚡

Agent Workflow & Pipeline Automation

DevOps & Automation Suite

CLI batch processors for SFX alignment, subtitle timing, auto-cleaning, and automated ComfyUI WebSocket queuing.

Scripts: pipeline.py • run.py Active
Python 3.12 • Llama.cpp

Installed Local AI Models (D:\ ... \local_ai\)

Optimized for NVIDIA GeForce RTX 5070 (12GB VRAM) • 32GB RAM

Open ComfyUI Studio (:8188) ↗ Open Llama-Server (:8081) ↗
🎬 VIDEO GENERATION MODELS
MiniMax-H3 FL2VA Pruned (FP8 Scaled) 20.96 GB • Video + Native Stereo Audio • FL2V Interpolation
Ready
Wan 2.1 Image-to-Video 14B (Q3_K_S) 7.38 GB • High-fidelity physical motion • 832x480 (16fps)
Ready
MiniMax-H3 Turbo 4-Step LoRA 1.82 GB • 4-Step Ultra-Fast Sampling
LoRA
🎨 IMAGE & CHECKPOINT MODELS
Illustrious-XL v2.0 (SDXL Anime Official) 6.46 GB • Anime 2D/2.5D & Dark Fantasy Masterpiece
Ready
CyberRealistic XL Play v10.0 (FP16) 6.46 GB • Photorealism, Studio Lighting & Portraiture
Ready
PonyRealism v2.2 (Main VAE) 6.62 GB • Anatomical Precision & Complex Posing
Ready
FLUX.2 Klein 4B + Lustly LoRA 4.1 GB • DiT Transformer Architecture
Ready
🕹️ CONTROLNET & POSE GUIDANCE
xinsir OpenPose SDXL 1.0 2.33 GB • Full Body Skeleton & Pose Locking
Ready
xinsir Canny Edge SDXL 1.0 2.33 GB • Contour, Armor & Costume Consistency
Ready
🧠 LOCAL LLMs (LLAMA-SERVER :8081)
Qwen2.5-Coder-14B-Instruct (Q4_K_M) 8.99 GB • Coding Specialist, Pipeline Debugging, Continue IDE
Active on :8081
Qwen2.5-14B-Instruct-Abliterated (Q4_K_M) 8.99 GB • Uncensored Scripting, Storytelling & Creative Brainstorming
Standby
🟢 Llama-Server Ready at 127.0.0.1:8081
Qwen 2.5 Coder:

Hello Phong! I am connected to your local llama-server running on your RTX 5070 GPU. How can I assist with your movie pipeline, scripts, or coding tasks today?

Keyframe A (Start)
Drag image or select reference
Keyframe B (FL2V End)
Drag image or select reference
Estimated time: ~3.5 minutes on local GPU
EPISODE 01

Call of the Forest

03:15 • 14 Scenes • 960x960 24FPS

Hardsub Master Ready
EPISODE 02

Void Fracture

04:02 • 18 Scenes • 960x960 24FPS

Hardsub Master Ready
EPISODE 03

Primordial Flame

04:45 • 22 Scenes • 960x960 24FPS

Hardsub Master Ready
EPISODE 08

Eclipse Awakening (Final Arc)

Style: Dark Fantasy Impasto Oil Painting

Rendering MiniMax-H3 Scenes (80%)