We’re looking for an engineer who can work across Generative AI, model fine-tuning, optimization, and on-device deployment for iOS and Android.

This role is ideal for someone who enjoys taking models beyond experimentation and making them actually work on consumer devices under real constraints such as latency, memory, battery, thermal limits, and model size.

What you’ll work on

• Image Generation, VTON, identity-preserving and reference-based generation
• LoRA training, fine-tuning, personalization, and adapter-based techniques
• PyTorch-based training and experimentation
• Diffusion / Transformer-based image generation architectures
• Dataset preparation, training pipelines, and evaluation
• Model conversion and deployment using ONNX / ONNX Runtime
• Inference optimization using TensorRT
• Quantization using FP16 / INT8 / INT4 and other optimization techniques
• On-device inference on iOS using Core ML and Android using TFLite/LiteRT, ONNX Runtime, or similar runtimes
• Performance profiling and debugging across CPU, GPU, NPU / ANE
• Reducing inference latency, peak memory usage, and model footprint
• Building production-ready model pipelines from training to device deployment

What we’re looking for

• 2-5 Years of Experience
• Strong experience with Python, PyTorch, and deep learning
• Hands-on experience with Computer Vision / Generative AI
• Experience training or fine-tuning image-generation models
• Good understanding of LoRA, model optimization, and quantization
• Strong knowledge of inference runtimes such as ONNX, TensorRT, Core ML, TFLite/LiteRT, or ONNX Runtime
• Strong debugging, profiling, and software engineering fundamentals

Good to have

• Experience with FLUX, Stable Diffusion, SDXL, ControlNet, IP-Adapter, VAE, CLIP, or similar architectures
• CUDA / GPU optimization
• Metal / Apple Neural Engine
• Android NNAPI / GPU delegates
• C++ / Swift / Kotlin
• Distributed or multi-GPU training
• Model compression, pruning, and knowledge distillation

We’re especially interested in engineers who can think end-to-end:

Data → Training → Fine-tuning → Evaluation → Quantization → Model Conversion → Runtime Optimization → On-Device Deployment → Production Monitoring

If you enjoy solving problems like “How do we make a large generative model run fast, efficiently, and reliably on a phone?”, this role should be exciting.