I don't know YET, but I'm skilled at finding the answers.
🚀 Learn and Experiment with AI technology ðŸ§
LLM | Agents | MLOps | ML
-
LinkedIN |
GitHub |
Hugging Face |
Kaggle
Pet Projects
Recent Blog Posts
Here are some of my latest blog posts:
-
Transformer FFN with SwiGLU (October 09, 2025)
-
Rotary Position Embeddings (RoPE)- Intuition, Math, and Examples (October 09, 2025)
-
Architecting for Stability- Pre-Normalization, RMSNorm, and Mixed-Precision Training (October 08, 2025)
-
From tokenizer to uint16 dataset with encode_iterable (September 21, 2025)
-
10K vs 32K Tokenizers Yield Similar Bytes per Token (September 21, 2025)
-
Bytes → UTF-8 → BPE — why not just number the alphabet? (September 21, 2025)
-
Books List (March 25, 2024)
-
Visual Explanation of Transformer with Dimensions (March 25, 2024)
-
LLM Prompts Dump (March 17, 2024)
-
Attention Mechanism (March 17, 2024)
-
Speculative Inferencing (March 05, 2024)
-
LLM Inferencing Optimization (February 28, 2024)
-
Cheat Sheet- Understanding Kubernetes Architecture (February 01, 2024)
-
Understanding LLM Inferencing Challenges and Tools (January 29, 2024)
-
Mixture of Experts (MoE) (January 14, 2024)
-
KV Cache in Transformers- Detailed and Simplified Guide (January 03, 2024)
-
Quantisation LLM (December 22, 2023)
-
LLMOps 1 (November 01, 2023)
-
LLM buy vs build (November 01, 2023)
-
Python Production (November 01, 2023)
-
Setting Up WSL Terminal Windows (October 21, 2023)
-
Paper - ATTENTION SINKS (October 15, 2023)
-
Open Source Contribution (October 09, 2023)
-
Your Blog Post Title (October 09, 2023)
-
Paper - PROMPTBREEDER (October 09, 2023)
-
Welcome to Jekyll! (October 01, 2023)
Current Workplace
- Company Name: Fractal Analytics
- Position: AI Engineer
- Brief Summary:
- At Fractal Analytics, I’m at the forefront of AI, leading the development of Large Language Models (LLM) applications using retrieval augmented generation (RAG). From customized document generation to architecture development and code conversion, I’ve got it covered.
- MLOps is my playground where I streamline pipelines, contribute to data warehouse migration, and develop custom scripts to maintain and monitor the ML models in production.