Yashwanth Sai

Experience

1. Ai Engineer

   Stealth Startup

   May 2025 — Present · Los Angeles, CA, US

   Deployed and optimized AI model serving infrastructure supporting multiple model types using vLLM and PyTorch on GPUs, achieving 90% GPU utilization and high throughput inference for scalable customer applications.·Implemented a microservices based AI deployment pipeline using Docker containers and LiteLLM proxy routing, reducing deployment time by 60% and automating model orchestration across 20+ LLMs, improving inference reliability by 30% and enabling cost efficient scaling for customer applications.·Optimized AI model efficiency using LoRA/QLoRA quantization, reducing VRAM usage up to 50% and enabling deployment on cost effective GPUs, saving 30% on infrastructure costs, while fine-tuning models for multimodal tasks·Delivered multimodal AI capabilities including OCR, speech synthesis, and vision processing via SkyPilot’s intelligent orchestration to manage cost efficient infrastructure and enabling seamless integration for diverse customer use cases.

Education

• PES College of Engineering, Mandya

  Bachelor of Engineering - BE, Computer Science

• University of California, Riverside

  Masters in Computer engineering