Software Engineer
Applied AI leader building smart, fast software for big tech companies
Hello! I’m Akansha Bansal.
I work at the intersection of ML compilers and AI infrastructure, architecting systems that optimize machine learning workloads across hardware platforms. At AMD, my team develops the SHARK compiler infrastructure that powers AI model deployment for industry leaders including Tesla, Meta, and Google. I’m particularly drawn to the challenges of computational efficiency – from compiler optimizations to enterprise-scale distributed systems.
My engineering journey reflects the evolution of modern computing stacks. Starting with analog IP design for Intel’s microserver chips, I progressed to building enterprise infrastructure at Cisco, where I led WebEx’s 5X platform scaling during unprecedented growth. At TikTok, I architected video processing algorithms and generative AI tools that enhanced content creation for millions of users, while collaborating with Snapchat on inclusive camera features. Each role has deepened my expertise in building efficient, scalable systems across enterprise and consumer technologies.
WORK WITH ME
Looking for professional growth?
We’re here to help you unlock new opportunities and scale with confidence. Our tailored coaching are designed to meet your unique needs, driving your growth through innovative strategies and proven techniques.
Innovating brilliance
Everything we learned about DeepSeek V3
DeepSeek codes, ChatGPT talks better.
DeepSeek V3 and ChatGPT represent two distinct approaches to large language model development, with DeepSeek V3 featuring a 600B parameter open-weight architecture trained on 14.8T tokens, while ChatGPT operates on a proprietary closed architecture with a strong emphasis on safety and ethics. DeepSeek V3 distinguishes itself through superior performance in technical tasks, including code optimization and algorithm implementation, while offering cost-effective training and deployment options. Its open architecture enables customization and strong multilingual capabilities, making it particularly suitable for research and development projects, custom AI development, and specialized technical applications where resource efficiency is paramount.
In contrast, ChatGPT has established itself as a leader in natural language understanding and conversational abilities, excelling in content creation and cross-domain knowledge application. It offers robust safety measures, bias control, and enterprise-grade API infrastructure, making it ideal for business applications, customer service, and general-purpose AI solutions. ChatGPT’s managed infrastructure comes with professional support, regular updates, and an established pricing structure, though at a higher operational cost compared to DeepSeek V3. While both models have their strengths, ChatGPT focuses on enterprise solutions and content generation, whereas DeepSeek V3 caters more to technical and research-oriented applications.