Hafiz Muhammad umar Akram

Designing Deep Learning based algorithms to automate complex power systems. Designing/developing the complete pipeline from fetching the data from sensors to inference and displaying the outputs on grafana dashboard. Anomaly detection in the electric circuits.

Developed Machine Learning piplines, from data wrangling phase to model implementation phase. Clustering of unlabeled images into document types using unsupervised learning techniques after extracting image embeddings. Manages the complete project of automatic number plate recognition. Designed OCR for a government body. End to end development of multiple reporting dashboards Developed data warehousing pipeline for real time reporting. Exploratory data analysis of users to enhance everyday performance

We designed an application whose job was to pick a website containing the desired customer/user products on the basis of their prices and reviews. Initially, we scrapped the products (reviews+price+information about the product) from multiple Pakistani websites and then we used ML techniques to categorize the reviews into good and bad ones. As a result, it helps our users to find the best platform to buy their desired products.

Machine Learning, Computer Vision

Information Processing Techniques, Concurrent and Distributed Systems, Data Mining, Deep Learning, Computer Networks, Probability, Linear Algebra