Syeda Yumna Zaidi Teacher Assistant at George Mason University

A

Data Scientist

Actfore

Jan 2024 - Current (2 years 6 months)

Engineered production-grade data extraction pipeline leveraging LLM models (Qwen, Llama, Gemma) and ROLMOCR vision model to process unstructured documents at scale, extracting sensitive PHI/PII data with 95%+ accuracy • Built comprehensive data scraping workflow utilizing NLP, regex, image processing, and OCR for named entity recognition and pattern matching, processing 1M+ files within hours using parallel processing and CUDA optimization • Developed Python-based B2B client search application with automated ETL pipelines, handling diverse file types from database dumps to text files, enabling rapid compliance reporting for regulatory requirements • Designed ML-based filename matching algorithm using exact and fuzzy matching techniques to i

Teacher Assistant

GEORGE MASON UNIVERSITY

Aug 2023 - Dec 2023 (5 months)

Assisting in the HCI Course, providing support and guidance to students.

Program Analyst Intern

USDA

Jun 2023 - Aug 2023 (3 months)

- Gathering relevant IT spend data from various sources within the Agriculture Department, including financial records, procurement documents, IT applications portal and budget reports. - Performing comprehensive analysis of the IT spend data to identify patterns, trends, and outliers. This also involves using statistical methods, data visualization techniques, and comparative analysis to gain insights into spending patterns. - Evaluate the efficiency and effectiveness of IT spending to identify opportunities for cost optimization. It also includes finding applications which are not effective for optimizing agricultural operations.

Data Analytics Intern

Global Business Alliance

Feb 2023 - May 2023 (4 months)

- Conducting research and creating statistical models to advance GBA’s advocacy and membership initiatives. - Gathering and analyzing data on elections and federal legislative records using excel and tableau. - Creating tableau hyper data extracts after cleaning and processing the collected data. - Generating statistical analysis on political and public policy trends, as well as membership engagement, including predictive analytical tools. - Creating tableau dashboards and reports for internal and external audiences.

Research And Development Scientist

Love For Data

Jun 2019 - Dec 2021 (2 years 7 months)

- Analyzed surveillance videos from all around Pakistan to rate a location's compliance with COVID SOPs, using Computer Vision models. Government of Pakistan used them for implementing Smart Lockdowns. Deep learning models were used for Face Mask detection and Human Detection to identify the distance between individuals. For each video the weighted sum of average Naked Face Count, Close Human interactions Count, and Population were used to get a scale from 0 to 10 for each location. Locations with high scales were the most vulnerable. The solution was built in Python. Keras and Pytorch were used for deep models. The work was acknowledged internationally. - Trained a deep learning model to remove artifacts from bank cheque images. Tensorflow

Data Scientist

Habib Bank Limited

Oct 2017 - Jun 2019 (1 year 9 months)

- Developed ETLs using SSIS and Alteryx to combine core-banking databases with other banking databases to build a centralized Data Mart for Analytics. ETLs were written using SQL queries getting data from multiple servers. These ETLs were executed at night. The ETLs were dumping data to an RDMS established on a local server using SQL server. - Built analytical dashboards using Python and Tableau, providing useful insights for Credit, Compliance, Loan, Consumer, Marketing, ATM Monitoring, and other departments. These dashboards were built with different levels of accessibility for different logins. The dashboards were reading data from a SQL database using a complex SQL query. - Anti-Money laundering dashboard was built using Python as it w

Q

Natural Language Processing Lead

Quarrio

Aug 2016 - Sep 2017 (1 year 2 months)

- Led a team of Five NLP developers involved in the development of NLP module. This module can convert text questions to database queries for retrieving data against a text question. - Developed the first version of the text processing module which can convert text question words to Named Entities, Verbs, Adjectives, and other tags. - Developed the pipeline which can transform words and their tags to SQL Query words in order to eventually formulate a query. This pipeline used pre-defined dictionaries for getting the query word against a detected tag. The words also have types described in the dictionaries. These fetched words and their types were used to formulate the query by defined rules in the code. - Trained a classifier using Stanfor

S

Data Science Trainer

Software Park Thailand

Feb 2016

Conducted a 7-day workshop on Data Science to a class of 25 professionals at Software Park, while working at PredictifyMe. Taught them major Statistical Concepts, Machine Learning Algorithms, Data Partitioning, Model Learning and Evaluation techniques.

P

Data Scientist

PredictifyMe

Jan 2015 - May 2016 (1 year 5 months)

Analyzed client’s data through the complete data cycle, - Analyzed Voxent’s data to identify patterns in their data. This analysis was done in R which included processing data to clean and normalize it, analyzing similar patterns by plotting various plots using ggplot2 for different groups and locations. Finally suggestions were also made based on the analysis which guided them to target critical data groups. - Performed Regression analysis for Coldwellbanker to identify the KPIs affecting their listings. The process included extracting data through SQL Queries after getting remote access to their server. The analysis was done in R to train a precise model which was later used to identify features most important for affecting listings. Age

G

Web Developer

Grappetite

Dec 2012 - Sep 2013 (10 months)

- Developed user interface of web applications using Html, Javascript, Jquery and CSS. - Used MVC framework for developing the complete architecture of the applications. - Established RDMS using SQL Server on the server. Implemented CRUD operations in the admin panel of the websites. - Developed PHP based back-end controller and model functions routing users to all the views and their interactions. Projects: mijnzorg, halat-o-meter, onderbouwd, studentperspectives.

P

Internee

Pakistan Steel Mills Corporation

Dec 2011

Worked at Steel manufacturing and plant controlling departments like HSM (Hot Strip Mill), Coke Oven, Blast Furnace, Electronic Laboratories etc.

Syeda Yumna Zaidi

About

Detail

Timeline

Résumé