Project Timeline
A chronological journey through my development milestones and technical achievements.
March 2025
Row-Tiered Mixed-Precision Quantisation for LLMs
Designed & implemented Variable Bit Floating-Point (vBFP) quantisation on Llama-3-8B.
C++CUDAPyTorch+1
Quantisation Algorithm
CUDA Kernel
Evaluation
November 2024
Extractify: Algorithmic Document Extraction & Visual Mapping IDE
Engineered a 5-step Visual ETL pipeline for heavily unstructured PDFs, eliminating manual data entry.
PythonReactFlask+1
Visual ETL Pipeline
N-gram Algorithm
Anomaly Engine
July 2024
GitChat: Context-Aware Developer AI Assistant
Developed a full-stack AI assistant for source-code repositories using Retrieval-Augmented Generation (RAG).
TypeScriptNode.jsReact+1
RAG Implementation
Context Management
Frontend Integration
April 2024
llm.karanmertiya.me
An ambitious project involving the deployment and serving of large language models.
PythonPyTorchLLMs+1
Model Optimization
Infrastructure Setup
Deployment
October 2023
Pixie Net
A deployed vision model for image processing and classification tasks.
PythonComputer VisionPyTorch
Model Training
Evaluation
Production Deployment