Ajith Prakash Sirra

About Ajith Prakash Sirra

Experienced AI/ML Solution Architect with 8+ years in building scalable AI solutions across NLP and Computer Vision domains. Skilled in optimizing LLM workloads, conducting performance benchmarking, and developing inference pipelines on platforms like AMD, Intel, and Nvidia. Proficient in deep learning frameworks (PyTorch, TensorFlow) and architectures such as GPT, LLaMA, and BERT. Known for delivering impactful technical insights, contributing to industry conferences, and advancing AI infrastructure through innovative system-level solutions.Key Skills: Deep Learning & AI Optimization – Extensive experience in training and optimizing Deep Neural Networks (DNNs) for Computer Vision and NLP tasks, ensuring efficient inference and deployment. Hardware-Accelerated AI – Proven track record in accelerating end-to-end AI solutions on NVIDIA and AMD-Xilinx hardware platforms for high-performance deployment. End-to-End AI/ML Pipelines – Skilled in designing and implementing scalable AI/ML pipelines for Vision and NLP applications, from prototyping to production. AI Performance & Cost Analysis – Expertise in benchmarking AI model performance across hardware architectures and conducting Total Cost of Ownership (TCO) analysis for optimized AI solutions. AI Market & Trends – In-depth understanding of the AI industry landscape, emerging technologies, and future advancements.Frameworks Familiar with : PyTorch Tensorflow Keras ONNX TensorRT-LLM VITIS-AI OpenVINOCore Competencies : LLaMA Mixtral MoE GPT DeepSeek Grok BERT LSTM Models CNN Models

Experience

1. Mts Product Application Engineer - Ai Ml

   AMD

   Jul 2023 — Present · Hyderabad, IN

   AI Solutions Professional at AMD | 6+ years in Strategic AI Solutions Group (AIG)| Specializing in hardware-optimized AI systemsKey Responsibilities : • Leading competitive analysis of SOTA LLM workloads on AMD vs Intel host platforms,showcasing AMD\'s 20% higher performance in GPU-accelerated Training and Inference.• Creating technical white papers showcasing AMD\'s performance advantages, deliveringinsights through data-driven benchmarking and summary reports.[Report on AMD EPYC]• Showcasing 10% TCO benefit for NLP Solutions on AMD\'s vs competitor platforms.• Creating Inference Pipelines of Vision and NLP workloads on AMD’s AI Software Stack.• Evaluating LLM performance on AMD vs Nvidia GPUs for competitive analysis BM reports.• Analyzing LLM workload execution, identifying key hotspots and suggesting optimizations.• Performance projection of LLM performance on current and upcoming Xilinx Accelerators.• Collaborated on AMD\'s Vitis-AI Software Stack, addressing current and future roadmaps.• Optimized AMD’s Vitis-AI Software Stack for improved User and Developer Experience.

Education

• Visvesvaraya National Institute of Technology

  Master’s Degree, Communication Systems Engineering

  2014 — 2016

• SRKR Engineering College

  Bachelor’s Degree, Electronics and Communications Engineering

  2009 — 2013