About The Role
You will lead the design and deployment of:
- photorealistic AI fashion generation systems,
- GPU inference pipelines,
- SDXL-based multi-stage rendering,
- high-resolution upscaling,
- ethnicity-aware model systems,
- garment preservation architecture,
- and scalable AI infrastructure.
This is NOT a standard web engineering role.
We need someone deeply experienced in:
- Stable Diffusion ecosystems,
- GPU optimization,
- AI orchestration,
- model fine-tuning,
- and production-scale image generation systems.
Core ResponsibilitiesAI Fashion Generation Pipeline
Design and optimize:
- SDXL generation workflows
- multi-stage diffusion pipelines
- LoRA orchestration
- ControlNet pose conditioning
- IPAdapter garment preservation
- fashion relighting systems
- face consistency pipelines
- high-resolution image refinement
High-Fashion Image Quality
Implement:
- HDRI lighting systems
- studio lighting presets
- beauty dish simulation
- cinematic fashion relighting
- skin-tone accurate rendering
- luxury editorial aesthetics
- photorealistic skin textures
- fabric detail preservation
Garment Preservation System
Build and optimize:
- segmentation pipelines using SAM2
- garment detection using GroundingDINO
- IPAdapter conditioning
- ControlNet Tile workflows
- texture-preserving diffusion refinement
Goal:Maintain accurate garment fit, texture, logos, folds, and patterns across generations.
Ethnicity & Diversity Modeling
Develop:
- ethnicity-specific LoRAs
- skin tone calibration systems
- demographic-aware relighting
- weighted model pools
- inclusive AI rendering systems
Must understand:
- color science,
- histogram balancing,
- CIELAB calibration,
- and realistic multicultural skin rendering.
High-Resolution Rendering Pipeline
Implement:
Generate 1024
Refine 1536
Upscale 2048
Tile Upscale 4096
Face Restore
Sharpen Fabric
Using:
- SDXL latent upscale
- RealESRGAN_x4plus
- Tile Diffusion
- GFPGAN
- CodeFormer
Pose & Motion Systems
Build:
- editorial pose libraries
- runway poses
- ecommerce poses
- lifestyle poses
- fashion motion systems
Using:
- OpenPose
- DW Pose
- MediaPipe
- ControlNet
AI Infrastructure & MLOps
Own:
- GPU orchestration
- Dockerized inference workers
- CI/CD for AI pipelines
- model deployment/versioning
- autoscaling GPU systems
- inference optimization
- queue architecture
- observability & monitoring
Required Technical SkillsMUST HAVEGenerative AI
- Stable Diffusion XL (SDXL)
- LoRA training
- DreamBooth
- Diffusers
- ControlNet
- IPAdapter
- InstantID / PhotoMaker
- ComfyUI workflows
Computer Vision
- OpenCV
- segmentation systems
- pose estimation
- image enhancement
- relighting pipelines
Fashion AI Systems
Experience with:
- fashion photography generation
- garment preservation
- editorial image generation
- high-resolution AI upscaling
- photoreal human rendering
AI Infrastructure
- Python
- FastAPI
- PyTorch
- CUDA
- GPU optimization
- TensorRT (bonus)
- Triton Inference Server (bonus)
DevOps / Platform
- Docker
- Kubernetes / EKS
- Redis / BullMQ
- AWS GPU infrastructure
- S3 / CloudFront
- GitHub Actions CI/CD
- Prometheus / Grafana
- Sentry
Strong Bonus Experience
- AI ecommerce systems
- Virtual try-on pipelines
- Human image synthesis
- AI relighting
- 3D fashion pipelines
- ComfyUI custom nodes
- RunPod / LambdaLabs / Modal
- Multi-GPU orchestration
- AI inference optimization
- Real-time generation systems
AI Stack We Use
Component
Stack
Base Models
SDXL
Pose
ControlNet
Identity
InstantID
Garment Preservation
IPAdapter
Segmentation
SAM2
Upscaling
RealESRGAN
Face Restoration
GFPGAN / CodeFormer
Backend
FastAPI + NestJS
Frontend
Next.js
Queue
BullMQ
Infra
AWS
CDN
CloudFront
Storage
S3 / R2
Monitoring
Prometheus + Grafana
Key Problems You Will Solve
- Maintaining garment fidelity
- Achieving luxury-grade realism
- Consistent face identity
- Accurate skin-tone rendering
- High-quality 4K generation
- GPU cost optimization
- Inference speed optimization
- Multi-ethnicity realism
- Fashion-quality lighting systems
- Production AI orchestration
Ideal Candidate
You are someone who:
- understands both AI research and production infrastructure,
- has built real-world diffusion systems,
- can optimize GPU-heavy pipelines,
- understands visual quality deeply,
- and can scale AI systems reliably.
This role requires both:
- engineering depth,
- and aesthetic sensitivity.
Remote / Contract / Full-Time
Open to:
- Remote
- Contract
- Fractional senior consultant
- Full-time lead engineer
To Apply
Please include:
- LinkedIn/GitHub
- Portfolio or demos
- Relevant AI fashion / diffusion work
- GPU infrastructure experience
- SDXL / LoRA / ControlNet projects
- Any fashion-tech or image-generation experience
Pay: From ₹25,000.00 per month
Work Location: Remote
