Job Title

Location:Remote (India-based preferred) Type:Full-time | Founding Team | High Equity Company:Flickd ()About the Role We’re buildingIndia’s most advanced virtual try-on engine— think Doji meets TryOnDiffusion, but optimized for real-world speed, fashion, and body diversity.As ourML Engineer (Computer Vision + Try-On) , you’llown the end-to-end pipeline : from preprocessing user/product images to generating hyper-realistic try-on results with preserved pose, skin, texture, and identity.You’ll have full autonomy to build, experiment, and ship — working directly with React, Spring Boot, DevOps, and design folks already in place.This is not a junior researcher role. This isone person building the brain of the system- and setting the foundation for India's biggest visual shopping innovation. What You’ll BuildStage 1: User Image Preprocessing Human parsing (face, body, hair), pose detection, face/limb alignment Auto orientation, canvas resizing, brightness/contrast normalization Stage 2: Product Image Processing Background removal, garment segmentation (SAM/U^2-Net/YOLOv8) Handle occlusions, transparent clothes, long sleeves, etc. Stage 3: Try-On Engine Implement and iterate on CP-VTON / TryOnDiffusion / FlowNet Fine-tune on custom data for realism, garment drape, identity retention Inference Optimisation TorchScript / ONNX, batching, inference latency minimization Collaborate with DevOps for Lambda/EC2 + GPU deployment Postprocessing Alpha blending, edge smoothing, fake shadows, cloth-body warpsYou’re a Fit If You:Have2–5 yearsin ML/CV with real shipped work (not just notebooks) Have worked on:human parsing, pose estimation, cloth warping, GANs Arehands-on with PyTorch , OpenCV, Segmentation Models, Flow or ViT Can replicate models from arXiv fast, and care about output quality Want toown a system seen by millions , not just improve metricsStack You’ll UsePyTorch, ONNX, TorchScript, Hugging Face DensePose, OpenPose, Segment Anything, Diffusion Models Docker, Redis, AWS Lambda, S3 (infra is already set up) MLflow or DVC (can be implemented from scratch)For exceptional talent, we’re flexible on cash vs equity split. Why This Is a Rare OpportunityBuild the core AI productthat powers a breakout consumer app Work in azero BS, full-speed team(React, SpringBoot, DevOps, Design all in place) Bethe founding ML brainand shape all future hires Ship in weeks, not quarters — and see your output in front of users instantlyApply now, or DM Dheekshith (Founder) on LinkedIn with your GitHub or project links. Let’s build something India’s never seen before.