About Me

I'm passionate about Machine Learning and AI, with experience in deep learning and computer vision. In my free time, I enjoy exploring new AI technologies like NLP and generative AI.

I'm currently a Research Assistant at the University of Auckland, where I work with generative diffusion models to generate synthetic neuronal histology data and reduce the need for manual annotation.

I'm looking for opportunities to apply my skills and passion in Machine Learning and AI, while continuing to learn and build meaningful projects.

Feel free to check out my resume!

Resume

My Tech Stack

Technologies

PyTorch

Hugging Face

scikit-learn

NumPy

Matplotlib

Pandas

React

FastAPI

LangChain

HTML

CSS

Languages

Python

JavaScript

R

MATLAB

SQL

Here's some of my projects

• 

  Diffusion-Based Data Generation for Neuronal Instance Segmentation

  Python, PyTorch, Matplotlib, scikit-image, OpenCV

  My current research project exploring diffusion models for generating synthetic neuronal histology image-mask pairs for instance segmentation. Developed a two-stage conditional diffusion pipeline to generate paired images and segmentation masks without manual annotations, with the goal of improving instance segmentation datasets.
  
  This project has strengthened my skills in generative AI, PyTorch, and computer vision while providing hands-on research experience.

• 

  Research Paper Query Assistant

  Python, Streamlit, LangChain, LangSmith, Pinecone, Groq, Hugging Face

  I built this project to make it easier to understand technical research papers by allowing users to ask questions directly against their own documents. The system uses a RAG pipeline to retrieve relevant information from documents using Hugging Face embeddings and Pinecone, then provides that context to an LLM through Groq to generate answers. The entire pipeline was built with LangChain and presented through an interactive Streamlit UI, with a focus on improving factuality and reducing hallucinations.
  
  I developed my understanding of RAG, embeddings, vector databases, and LLMs while learning how to use LangChain, Pinecone, Hugging Face, and Groq. It was a challenging project, but it was worth the learning experience
• 

  Digit Classifier Project

  Python, PyTorch, FastAPI, Docker, React, JavaScript, HTML, CSS

  My first computer vision project, where I trained a Convolutional Neural Network (CNN) on the MNIST dataset using PyTorch to classify handwritten digits. I then deployed the model through a full-stack web application using React, JavaScript, HTML, CSS, FastAPI, and Docker.
  
  This project was both challenging and fun, helping me develop my foundational deep learning skills with PyTorch while also giving me hands-on experience with model deployment. The web application was deployed using Render.