Muhammad Abrar Khalid

As a Fintech Data Scientist, I have expertise in data technologies like Python, Deep Learning, SQL, and Scala. My work helped clients by generating reports to analyze customer transactions, identifying key features for an AI fraud prevention model, and reducing fraud losses by $150,000. I also improved Deep Learning efficiency by 60X through distributed GPU computing with Horovod and developed a customized fraud prevention model.

I developed a semantic search application using NLP techniques, aimed at reducing assessment time for lawyers in the insurance industry. I extracted KPIs using NER and employed LSH to find similar accidents. The project utilized Python, PyTorch, TensorFlow, FastAPI, HuggingFace, and Docker.

Project Goal: Identify problem statement from the literature in the field of graph mining. Solution: Proposed a methodology to reduce the reconstruction error from a summarized graph, We proposed a pattern based corrections methodology to maintain the only important corrections. Result: We were able to achieve less reconstruction error(false positive edges) rate from a summarized graph with the help of pattern based corrections. Tools: Python, MATLAB, NetworkX, iGraph, Pandas, Numpy.

As a BI Consultant, I designed complete BI semantic layers, developed SQL procedures, and met deadlines while achieving client expectations. I worked with multiple BI tools like Tableau, Power BI, and MicroStrategy to improve business performance.

Worked on the development, support, maintenance and implementation of a complex project module. Performed technical scoping studies, requirement gathering and suggesting technical solutions. Worked under pressure and able to perform demo, project documentation and implementation.

Build apps and demos. Demonstrate the newest technologies and host tech events on our campus. Acquire the tools and training to lead technology discussion on our campus. Build apps