Ayush Thakur
Note: This website was created using Cursor and Claude 3.7 Sonnet. It’s currently a work in progress and may contain errors or inaccuracies. I apologize for any incorrect attributions or information. I’m continuously improving the site to ensure everything is accurate and up-to-date.
Ayush Thakur
Machine Learning Engineer at Weights & Biases
Google Developer Expert Kaggle Notebooks Master
Hello! I’m a Machine Learning Engineer specializing in LLM applications, model evaluation, and MLOps. I lead the AI team in India and lead the OS integration efforts. My work focuses on baking MLOps pipelines for open-source repositories, building W&B Weave enabled tracing and monitoring, evaluating LLM systems, and developing educational content for the open-source community. Most of my work can be found in the Authoring section.
I have developed MLOps solutions for popular repositories like Keras, OpenMMLab, few from Meta and more. I have also built integrations with OpenAI (fine-tuning and chat completions), Cohere, CrewAI and more.
These days I am exploring agents and MCPs.
I’m also a Google Developer Expert in Machine Learning and a Kaggle Notebooks Master.
When I’m not coding, I enjoy watching anime (Naruto is my all-time favorite!), sharing knowledge through technical articles, and collaborating on interesting ML projects.
Featured Work
RAG Techniques: From Naive to Advanced
A comprehensive guide to Retrieval Augmented Generation techniques for enhancing LLMs with external knowledge sources.
LLM Evaluation Framework
A systematic approach to evaluating LLM performance using W&B Sweeps.
SwAV-TF
TensorFlow implementation of self-supervised visual representation learning through contrasting cluster assignments.
Areas of Expertise
LLM Applications
Building, evaluating, and optimizing LLM-based applications.
RAG Systems
Implementing retrieval-augmented generation for knowledge-intensive applications.
MLOps
Developing robust ML pipelines and infrastructure for continuous delivery.
Computer Vision
Implementing and optimizing visual recognition systems.
Technical Writing
Creating accessible educational content on ML topics.
Model Evaluation
Systematically assessing ML model performance across dimensions.
Courses & Education
I’ve created comprehensive courses on: - RAG++: From POC to Production - Training and Fine-tuning Large Language Models (LLMs)