Raj Thakur — Senior Software Engineer

As a Machine Learning Engineer, I specialise in fine tuning LLMs, building high throughput inference serving platform that work with production workloads, with a deep focus on accelerating model performance on specialized ML Accelarators hardware. My expertise spans system design, distributed systems, and robust architecture, leveraging C++, Golang, and Java for scalable backends. I am particularly adept at driving performance gains in deep learning workflows, having extensively worked with VLLM, JAX and PyTorch/XLA to harness the power of AWS ML accelerators (Trainium/Inferentia). My contributions include disaggregated inference using vLLM and optimizing execution through PJRT integration and direct hardware interaction by developing the AWS Neuron SDK, ensuring maximum throughput and efficiency for complex models. I apply HLD, LLD, and design patterns to deliver high-performance, cost-effective ML solutions.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer with expertise in Fintech and Machine Learning.

Location: Santa Clara, California, United States

Experience: 11 yrs 4 mos