Description :

Role Overview :

We are seeking a highly experienced Senior Data Scientist specializing in Computer Vision and Generative AI to lead cutting-edge R&D and model deployment efforts.

You will drive innovation in vision-language and diffusion-based systems, from research prototypes to scalable production-grade pipelines.

Key Responsibilities :

- Design, train, and optimize deep learning models for image generation, image-to-image translation, and text-to-image synthesis.

- Conduct advanced R&D in Generative-AI, including Diffusion Models, Transformers, VAEs, and GANs.

- Explore Vision-Language and multi-modal learning (text, image, 3D) with models like CLIP, Flamingo, PixArt-, and Qwen2-VL.

- Implement distributed training and model parallelism using PyTorch, Deep Speed, and Hugging Face Accelerate.

- Lead experiments with LoRA, QLoRA, adapters, and parameter-efficient fine-tuning methods.

- Develop evaluation pipelines for visual realism, garment fidelity, and human perceptual quality metrics.

- Collaborate with product and engineering teams to deploy models to production using Triton, Torch Serve, or Fast-API.

- Design automated ML-Ops pipelines for training, evaluation, and deployment (CI/CD, monitoring, retraining).

- Mentor junior ML-engineers and data scientists, ensuring best practices in model reproducibility, experiment tracking, and code quality.

Required Skills & Experience :

- Hands-on experience in applied ML/AI, focusing on Computer Vision and Generative Models.

- Proficiency in image generation, in painting, style transfer, human parsing, and pose estimation.

- Hands-on experience with Stable Diffusion, ControlNet, LoRA fine-tuning, and Hugging Face Diffusers.

- Strong knowledge of architectures such as U-Net, ViT, Transformer-based encoders/decoders, VAE, and GAN variants.

- Expertise in distributed training, data parallelism, and GPU/TPU acceleration.

- Experience with synthetic data generation, texture alignment, and dataset curation pipelines.

- Proven track record of deploying production-grade ML systems (Fast-API, Triton, Torch Serve, or REST APIs).

- Proficient with Python, PyTorch, OpenCV, NumPy, Hugging Face, and WandB/MLflow