Hello!
I work on AI infrastructure at Bloomberg, building the systems that train and serve LLMs across Kubernetes and GPU clusters. When I get a chance, I write and talk about what goes on in the AI infrastructure space. I studied systems and ML at Carnegie Mellon University, and hold a bachelor's in Electrical Engineering from Delhi Technological University.
Posts
-
Securing and Scaling Jupyter at Bloomberg: How we run Notebooks on Kubernetes Talk
May 14, 2026
PyCon US 2026 · Long Beach, California
Spoke about how Bloomberg builds and scales its AI platform on Kubernetes, covering notebook orchestration, GPU management, AI-assisted workflows, and reliability measures.
-
Evaluating Tool Calling in Qwen3-4B: Base, Prompting and LoRA Blog
Feb 10, 2026
An experimental evaluation of Qwen3-4B Base fine-tuned with LoRA on structured, agentic tool-calling tasks.
-
Ensuring Reliability in a Cloud Native Jupyter Notebook Platform Talk
November 4, 2025
JupyterCon 2025 · San Diego, California
Spoke about Bloomberg's Jupyter Notebook platform, sharing insights on building and maintaining reliability in a cloud-native AI platform.
-
Model Caching for AI Workloads Talk
June 24, 2025
Open Source Summit 2025 · Denver, Colorado
Gave a talk on Model Caching for AI Training and Inference workloads to speed up model load time in a cloud native environment.
-
Quick thoughts on DSPy: automatic prompt optimization? Blog
Sep 8, 2024
A new paradigm for giving structure to prompt engineering
-
NNLIBC: Neural Networks in C for WebAssembly Project
June 1, 2022
This is a Pytorch-like neural network library written in C.
Contact Me
Email: rtjsingh30@gmail.com