Hansal Shah.
Transforming data into wisdom.
Graduate CS @ Cornell University
About Me
I am a Machine Learning Engineer with a passion for developing software solutions. I have a background in computer science and mathematics, and I have a strong interest in Generative AI & Natural Language Processing. I am currently a graduate student at Cornell in Ithaca, NY.
I am currently working on a RAG-based chatbot that will be used to provide vertical search results in the domain knowledge of a specific expert.
Here are a few technologies I've been working with recently:- Open Source LLMs
- Langchain
- PEFT - Huggingface models (LoRA/ QLoRA)
- Retreival Augmented Generation
- Pinecone
- React
- Fast APIs
- Python
Experience
- Managed data processing enhancements for real estate legal sales, handling complex JSONs and legal PDFs.
- Collaborated with the team to create an ELT Data Pipeline on Databricks using Python, DBT, Spark-SQL, and Elastic Search.
- Improved quantitative analytics and standardized real estate data using technologies like Delta Lakehouse Architecture and GPT.
- Led the development of a full-stack web application using ReactJS, enhancing business operations and user interaction
- Implemented an IoT-based feature in the application for real-time EV tracking on Google Maps, improving customer experience and vehicle management.
- Collaborated on back-end development with NodeJS, ExpressJS, and MongoDB, contributing to a more efficient and scalable system.
I interned as a Machine Learning Engineer at BISAG-N.
- Worked on extracting roads and paths from satellite imagery for military mapping purposes.
- Implemented an image segmentation solution with the U-Net model using Tensorflow & Keras.
- Achieved an Intersection over Union (IoU) score of 0.7, successfully enhancing military capabilities in mapping unknown regions.
Education
2023 - 2024
Master of Engineering in Computer Science
Cornell University
GPA:
4.02 out of 4.3
Extracurricular Activities
- Teaching Assistant - Designing Data Products (NBA 6070)
- Teaching Assistant - Machine Learning (CS 5780)
2019 - 2023
Bachelor of Technology in Computer Science & Engineering
Pandit Deendayal Energy University
GPA:
4.0 out of 4.0
Published four research papers in various journals and conferences:
- Cardiovascular disease prediction using various machine learning algorithms [Link]
- Enhanced handwritten digit recognition using optimally selected optimizer for an ANN [Link]
- Hand Gesture to Character Recognition using Convolutional Neural Network [Link]
- Integrated Database Management System for Emergency Services [Link]
Extracurricular Activities
- President of Brahmand - The Astronomy Club of PDEU
- Led a diverse team of 54 members across 6 departments.
- Planned and coordinated monthly events.
- Briefed department heads on their roles and event contributions.
- Oversaw financial and technical decisions made by the Treasurer, Vice President, and General Secretary.
- Acted as a mediator to resolve conflicts and ensure smooth operations.
Projects
RAG
PEFT
Guardrails
FastAPI
Python
Pinecone
Cloneadvisor - Chatbot
I led the development of a Retrieval-Augmented Generation pipeline, optimizing queries, implementing document re-ranking and fact-checking, reducing hallucination to 8.47%, and ensuring responses were accurate, non-controversial, and on-topic using robust guardrails.
React
Node
Bootstrap
Readme.md File Generator
Readme.md generator is a React application that takes basic information about the user's project as an input and generates a README file which the user can upload directly in their project's repository.
HTML
CSS
Javascript
Bootstrap
OS Virtual Lab
A web application that represents and runs different algorithms that Operating Systems use to run the computers. The application provides graphs for visualization, generates PDFs for explaining the math involved in calculations, and provides a way to compare the performance of different Disk Scheduling algorithms.
AI (Mini-Max)
HTML
CSS
Javascript
Tic-Tac-Toe
A web application provides a simple implementation of the popular game Tic-Tac-Toe. Here, you play against the computer and you have the freedom to make the first move and pick your playing symbol. The game implements the Mini-Max algorithm for the computer to choose an optimal move after you make your move. I dare you to win against the computer.
Tensorflow
Keras
ANNs
Python
EV-Navigator
A method for predicting the State of Charge (SOC), the State of Energy (SOE), and ultimately the range prediction using a sequence of neural network models. The technique predicts the range of the vehicle by taking into account the state of the battery, the state of the vehicle, the driving style, and the road conditions. The system is deployed as a website to visualize it's working in real life.
Achievements
Academic Gold Medalist
First Merit Rank - B.Tech CSE
Get in Touch
My inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!