Hamid Rezaee Email and Phone Number
I am Hamid (ˈhaː.med), and I study Information Science and Entrepreneurship as an undergraduate at the College of Arts and Sciences at Cornell University, class of 2026.Before I start talking about my professional journey, a bit about me: I'm a creative writer who's recently gotten into filmmaking. When I'm not deep in machine learning, you'll find me gaming - 'Need For Speed Unbound' is my current obsession. I'm super extroverted, so don't hesitate to reach out and chat about machine learning, LLMs, recent tech advancements, or just to discuss how adorable dogs are!Now, about my experiences:At Pequity in Ithaca, NY, I work as a Machine Learning Research Assistant. Here, I've accelerated an API-based data retrieval process by 3.6 times using multi-threaded Python implementation. I perform extensive data cleaning with Pandas, execute sentence embedding using SBERT, and enhance embedding clustering through K-Means and HDBSCAN.At Cornell Tech in New York City, I'm a Machine Learning Research Assistant working under Prof. Sabuncu, Vice Chair of AI Research at Cornell. I've engineered a loss function for accurately identifying cancerous and tumorous cells in brain MRI images. I've accelerated training efficiency using TorchDynamo APIs and implemented optimizers like Adamax to ensure training stability.With Cornell University Sustainable Design, I serve as the Machine Learning Sub-team Lead. I lead a team of five students in developing a model to predict HVAC components for energy efficiency. Through rigorous statistical analysis and feature engineering, we're working to optimize temperature calculation models.At LessonLoop, I was a Machine Learning Intern where I pioneered a Retrieval Augmented Generation (RAG)-based chatbot. I architected testing infrastructure on AWS Cloud9 and conducted comprehensive A/B testing of Bedrock LLMs, collaborating with AWS solution architects to evaluate performance.Drop me a message on LinkedIn or email me at hr328@cornell.edu if you want to connect!
-
Machine Learning Infrastructure EngineerE3 GroupIthaca, Ny, Us
-
Chief Technology OfficerLuga Ai Jan 2024 - PresentNew York, New York, United StatesAs a Chief Technology Officer at Luga AI, Inc., I am in charge of managing a team of 3 members consisting of a frontend engineer, a full-stack engineer, and a UX/UI designer. We are actively working to address the underrepresentation of Arabic language in generative AI space, making sure that services such as text generation, audio generation, and video generation are easily accessible to Arabic speaking community, mainly targeting communities based in Saudi Arabia. We are aiming to provide enterprise level, AI-powered solutions to public and private sectors in Saudi Arabia, enabling them to better serve the community. My main responsibilities include: - Hiring developers with the skillset necessary for pushing the project forward. - Managing the MVP development and providing clear instructions and feedback to the developers and designer. - Managing the cloud services and access management used for the development.- Supervising the code written by developers and verifying correctness and consistency.- Communicating the technical aspect of the startup to the potential investors and clients. - Keeping the technical team motivated by offering milestone-based rewards. -
Machine Learning Research AssistantPequity Sep 2024 - PresentIthaca, New York, United StatesAs a Machine Learning Research Assistant at Pequity, I focused on optimizing data processing pipelines and implementing advanced NLP techniques for improved data analysis. The techniques include extensive data filtering to get the optimal embedding, using Hugging Face's open-source embedding models, clustering the text embedding based on semantic similarity, and finally performing grid search to find the best configuration of the clustering. My responsibilities mainly include:- Optimizing API-based data retrieval process achieving 3.6x speed improvement through multi-threaded implementation in Python.- Managing extensive data cleaning and preprocessing operations using Pandas to prepare data for analysis.- Implementing sentence embedding techniques utilizing SBERT and Stella EN 1.5B models with focus on memory efficiency.- Developing and fine-tuning clustering algorithms using KMeans and HDBSCAN with comprehensive parameter grid search for optimal embedding clustering. -
Machine Learning Sub-Team LeadCornell University Sustainable Design (Cusd) Aug 2024 - PresentIthaca, New York, United StatesAs the leader of the machine learning sub-team, I am in charge of leading a team of 5 machine learning team members in developing a multi-variable regression model designed for predicting the time-difference value for a temperature change based on the user's room occupancy.My responsibilities mainly include:- Documenting robust statistical rational on predictions and running analysis to find correlations between variables for a careful selection of dimensions.- Experimenting with multiple machine learning algorithms and evaluating their robustness and generalization capabilities on noisy data.- Delivering a trained, production ready model with extensive error handling to the backend sub-team for a safe deployment. -
Machine Learning EngineerCornell University Sustainable Design (Cusd) Aug 2023 - PresentIthaca, New York, United StatesAs a member of the Cornell University Sustainable Design project team, I have built a LightGBM model that predicts the time when the heating or cooling systems should turn on and off to reduce energy waste. The prediction is based on the occupancy of the room as well as the time it takes for the room to get to the desired temperature. The model is trained on the data we obtained from building thermostats and room occupants regarding the operating hours of the heating and cooling systems. This model is planned to be deployed and moved to production on Fall 2024. In this project, I was focused towards:- Experimenting with different predictive models including linear regression from Statsmodels, Random Forest from Scikit-learn, and LightGBM. - Modeling the math behind the temperature difference calculation and predicting delta time.- Simulating a production-level testing by incorporating n data points iteratively and evaluating the performance improvement gradually. - Saving the model parameters to the local device as a binary file for device memory optimization and faster loading time. - Visualizing the model improvement using matplotlib for better interpretation and storytelling.
-
Machine Learning Research AssistantCornell Tech May 2024 - Aug 2024New York, New York, United StatesAs a machine learning researcher at Cornell Tech under Prof. Mert Sabuncu, Vice Chair of AI and Engineering Research at Cornell University, I have worked on creating a robust loss function that could significantly outperforms the current loss functions used in medical imaging, specifically on small lesion segmentation. My work consisted of training the model on the loss function and performing comprehensive hyperparameter optimization. My work mainly revolved around:- Developing M-Loss (multi-component loss) function and optimizing it using PyTorch.- Hyperparameter optimization and regularization of the model parameters to maximize generalization capabilities of the model on different datasets. - Managing and allocating GPU resources for training a UNet model for testing.- Parallelization of the processes for faster and more efficient training. - Interpreting the model output and logs and reporting it to the supervisor. -
Machine Learning InternLessonloop Aug 2023 - May 2024Chappaqua, New York, United StatesAs a Machine Learning Intern at LessonLoop, I collaborated closely with the CTO and CEO on product ideas that leveraged the capabilities of ML and generative AI for integration into our ecosystem. This included communicating with AWS solution architects and cloud specialists on the optimal testing and performance evaluation approaches.My work focus was mainly:- Testing a Retrieval Augmented Generation (RAG)-based chatbot. - Deployment of the testing stack on an instance and monitored on AWS CloudFormation. - Performance evaluation of different vector databases including OpenSearch Serverless Vector Database and Aurora with inputted data. - A/B testing multiple Bedrock Large Language Models (LLMs), including Amazon Titan, Meta Llama 2 family, and Cohere Command models. - Utilizing RAG to determine the best-performing LLM in incorporating the retrieved data into its responses. -
Machine Learning InternKquika Inc May 2023 - Aug 2023Queens, New York, United StatesAs a freshman machine learning intern, my role mainly centered around learning the implementation of different models. As a freshman who was new into the field, I was tasked with learning different techniques and implementations in ML and learning different aspects of it like approaches in improving generalization, tackling overfitting or underfitting, regularization, and more. My experience included:- Taking online courses to learn about machine learning and its use cases.- Implementing models like linear and logistic regression, k-means, random forest and XGBoost to familiarize myself with the frameworks. - Visualizing the metrics and the prediction to improve my storytelling and technical communication skills in ML. - Reporting the progress back to the CTO and asking for further advice and recommendations.
Hamid Rezaee Education Details
-
Information Science
Frequently Asked Questions about Hamid Rezaee
What company does Hamid Rezaee work for?
Hamid Rezaee works for E3 Group
What is Hamid Rezaee's role at the current company?
Hamid Rezaee's current role is Machine Learning Infrastructure Engineer.
What schools did Hamid Rezaee attend?
Hamid Rezaee attended Cornell University.
Not the Hamid Rezaee you were looking for?
-
Hamid Rezaee
Iran -
-
-
Free Chrome Extension
Find emails, phones & company data instantly
Aero Online
Your AI prospecting assistant
Select data to include:
0 records × $0.02 per record
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial