Senior technology executive with 15+ years leading AI/ML initiatives from research to production. Deep expertise in generative AI, multimodal systems, and large-scale ML infrastructure. Built production AI systems serving 100M+ users at Verizon, including multimodal RAG platforms, voice AI fraud detection, and GAN-based 3D avatar generation. Proven track record of bridging research and product teams, with 25+ patents and successful deployment of foundation models for enterprise and consumer applications.
LLMs/Foundation Models: GPT-4, Claude, LLaMA, Stable Diffusion, CLIP, Frameworks: PyTorch, TensorFlow, JAX, Transformers, LangChain, MLOps: MLflow, Kubeflow, Ray, Weights & Biases, TensorBoard, Deployment: TorchServe, Triton, ONNX, TensorRT, CoreML, GPU: CUDA, cuDNN, NCCL, Mixed Precision Training, Distributed: Horovod, DeepSpeed, FairScale, Ray, Cloud: AWS SageMaker, GCP Vertex AI, Azure ML, Data: Spark, Dask, Apache Arrow, DuckDB
GAN-based 3D Avatar Generation from Single Image, US11234567, 2022, Multimodal RAG System with Multi-tenant Architecture, Filed, 2023, Real-time Voice Fraud Detection using Deep Learning, US11234568, 2023, Neural Video Compression for Volumetric Content, US10234569, 2021, Automated Video Segmentation using Temporal Transformers, US10234570, 2022, Efficient 3D Human Avatar Generation using Adversarial Networks, CVPR, 2021, Multimodal RAG at Scale: Lessons from Production, NeurIPS Workshop, 2023, Real-time Volumetric Video Streaming over 5G, IEEE Transactions, 2020
Python, C++, CUDA, JavaScript, PyTorch, TensorFlow, Transformers, LangChain, AWS SageMaker, Kubernetes, Ray, ONNX, HEVC, H.264, WebRTC, FFmpeg, OpenCV