Awais Naeem
I will build custom AI & machine learning solutions.
New to Contra
Awais is building their profile!
I built a multilingual AI chatbot that ingests YouTube videos, transcribes with Whisper, stores embeddings in Pinecone and delivers real-time answers with clickable timestamps.
Built with Python, AWS SageMaker, EC2, RDS, and Vue.js. Supports English, Chinese, modular knowledge bases and full admin control.
0
4
Developed a machine learning model to detect diabetes using the BRFSS 2015 Diabetes Health Indicators dataset.
The project involved data preprocessing, feature engineering, and training multiple models, including Logistic Regression, SVM, Random Forest, MLP and XGBoost.
Parameter tuning was performed using GridSearch and KFold cross-validation. Class imbalance was addressed with undersampling techniques.
Separate notebooks were created for preprocessing, training and evaluation, showcasing an end-to-end workflow for diabetes prediction.
0
9
Fine-tuned LLaMA for Diagnosis-Related Group (DRG) prediction using sequence classification on hospital discharge summaries.
The model processes MIMIC-IV clinical data, extracting brief hospital courses to classify them into standardized DRG codes.
Implemented LoRA-based fine-tuning on LLaMA-7B, optimizing training efficiency.
Integrated a Gradio interface for live inference, enabling seamless DRG prediction.
Used preprocessing pipelines to map multi-year DRG data into a unified version.
0
15
In this project, I conducted a comprehensive data analysis and visualization of a video game sales dataset.
The analysis involved exploring sales trends across various platforms, genres, and regions using Python libraries such as pandas, NumPy and Matplotlib.
The insights gained from this analysis help game publishers and platform developers make informed decisions, particularly highlighting the dominance of North America and Europe in sales and the potential of specific genres in Japan.
This project also emphasized the importance of data-driven strategies in the gaming industry.
1
0
23