Hi , I'm Jeffrey J Sam

A graduate student at Rice University

🥼 I’m currently working as an Graduate Research Assistant. My work involves data wrangling, exploratory data analysis and gaining inference from machine learning model behaviours
🌱 I’m currently learning about Smooth optimization techniques,Tools for Big Date, ML at Scale ,etc.
🔭 I’m interested in Probabilistic Algorithms and Machine Learning
🛣 My most recent prject is a Unsupervised Probabilistic model to estimate the count of various vehicles using Locality-Sensitive-Hashing-based-Sketches and Computer Vision
👨‍💻 All of my projects are available at https://github.com/Jeffrey-Joan
💬 Ask me about Probabilistic programming, Machine Learning and Computer Vision
📫 How to reach me [email protected]
📄 Know about my experiences https://drive.google.com/drive/folders/1BZd8SqNIM7nCI27PcvpmRdOvc0AHUDX3
⚡ Fun fact My favourite TV show is The Office(US)

Connect with me:

Languages and Tools:

jeffrey-joan's Projects

amazon_ml_challenge_2021

Text classification using Naive Bayes classifier and Linear SVM for an Amazon ML challenge

brain-tumor-detection

Brain Tumor Detection through Transfer Learning

captioning-of-remote-sensing-images

We use a transformer model to caption images captured by remote sensors

locality-sensitive-hashing-based-sketches

This project aims to provide an unsupervised lightweight solution to estimate the count of various different category of Vehicles. By implementing a novel Locality Sensitive Hashing based sketch.

mlopspython

MLOps using Azure ML Services and Azure DevOps

What if we only update the minimum hash among all k options in Count-Min Sketch(This would reduce the number of updates that are applied to the counters) and also compress the number of rows to 1 like bloom filter(this would increase collisions)? who would this trade-off work. To test this out we run several experiments

Recommend Projects