Dayananda V

March 2022 – Present | Bengaluru, Karnataka

• Designed and implemented scalable solutions to improve distributed AI training stability and performance across multi-node environments.
• Collaborated with framework and hardware teams to optimize performance, scalability, and memory efficiency for large-scale AI workloads, improving overall system capabilities.
• Optimized memory and resolved framework-level issues for multi-node distributed workloads on Habana Gaudi 2 and Gaudi 3 hardware, enhancing system reliability.
• Contributed to PyTorch distributed framework enablement and porting support for Habana Gaudi hardware platforms.
• Worked on distributed communication and framework optimization to bridge gaps between AI frameworks and Intel accelerator hardware.
• Integrated SGLang framework with DeepEP framework to enable efficient Mixture-of-Experts (MoE) solutions for Intel XPU hardware in multi-node scale-up configurations.

• Led development of lightweight AI inference frameworks for microcontroller and edge-device-based AI solutions with optimized memory and performance characteristics.
• Core member of Huawei AI framework and compiler optimization team contributing to multiple AI solutions deployed globally across Huawei device platforms.
• Enhanced Huawei MindSpore AI compiler framework through development of generic operator fusion and pre-compiler optimization solutions.
• Specialized in AI model optimization techniques including quantization, pruning, graph optimization, and operator fusion.
• Deployed advanced computer vision models, including dehaze, deblurring, reflection removal, and spam filtering on Huawei handset platforms, improving user experience.
• Contributed to TensorFlow and Apache TVM community solutions involving framework integration, kernel optimization, and performance tuning.
• Worked extensively on Android framework development, Bluetooth stack integration, and AOSP platform optimization.
• Led and mentored engineering teams while contributing as designer, developer, architect, and technical lead across multiple projects.
• Developed tailored mobile solutions for AT&T, TMO, Verizon, and other international customers, addressing specific operational needs.