Saad Atta

Building an end-to-end Machine Learning Pipeline for data preprocessing to model training and hyper parameter optimization.

●Experience working with AWS S3, MWAA, Sagemaker and Databricks.

●Model tracking and serving via MLflow

●Processing and transforming raw data images using opencv and scikit-learn.

●Working with transformer-based models and pytorch.

●https://www.bravurasolutions.com/unitedkingdom/ Flask/ Django Computer Vision AWS Postgres NLP Pandas Elastic Search S3 C+ + Pytorch LSTM MLflow Spark CNN Docker Airflow

Built a model to predict valuation of used cars and real-estate properties for banks. I built the data processing pipeline and deployed predictive model on AWS server with Flask backend. Before building a Machine Learning model, I provided the clients with detailed statistical analysis of data using graphs and interactive visualizations. I tested and compared different Machine Learning models and used cross validation techniques to estimate the model performance. I deployed the best performing model into the production environment.

●Built a facial recognition based atm cash withdrawal prototype for banks. I used object recognition to identify if a person performing transaction is wearing a mask or helmet and block the user from making transaction. When the face is visible, the application recognizes the person and only then allows them to make a transaction.

●Experience working with Pandas, scikit-learn, tensorflow, CNN, RNN, LSTM, Flask, Postgres, React Native, Node Js, AzureML and AWS.

●https://valustrat.com/

Developed a workplace attendance application based on face recognition. The model was trained on all staff members and keeps record of ins and out of each employee. Involved in all the steps starting from data annotation till the deployment of final model.

●Built a model for predicting battery health in solar grids based on consumption, charging capacity and life of a battery over time. I used LSTMs to predict the power consumption and discharging of the battery over next hour. I also built an android mobile application for the same client to monitor power usage and solar energy.

●Experience working with Pandas, tensorflow, scikit-learn, CNN , LSTM, scikit-image, R, mongodb, apache spark, hadoop, and docker.

●https://www.inboxbiz.com/