Abdul Rehman Ahmed Shaik
● Plano, TX ● 316-***-**** ● ad5q1v@r.postjobfree.com ● Linkedin ● GitHub
Professional Summary
Confident young professional with a passion for learning and development, eager to leverage education and training in the data science space to support the growth and success of a high-performing organization. Strong track record of setting effective goals and leading teams to achieve them, committed to continuous improvement and driving team success. Dedicated candidate with strong analytical and problem-solving skills, skilled in gathering and analyzing data with a keen eye for accuracy. A reliable leader with strong mentoring and supervisory abilities, flexible and adaptable to dynamic environments. Highly motivated to drive innovative solutions.
Education
Master of Science in Computer Science August 2022- May 2024
Wichita State University
Bachelor of Technology in Electronic Communication Engineering August 2018- May 2022
Jawaharlal Nehru Technological University
Skills
Software Languages: Python, SQL, R, MATLAB, C
Model Evaluation, Optimization, Extract, Clean and Analyze Data
Technical Skills: Deep Learning, Machine Learning, GIT
Natural Language Processing Techniques: NLP, NLTK, spaCy
AI/ML Frameworks: TensorFlow, Keras, Scikit-learn, PyTorch, CNNs, RNNs, GANs, ResNet, LSTM, UNet
Data Visualization: Matplotlib, Seaborn, Tableau, pandas, Power BI
Database Management: MySQL, PostgreSQL, EC2
Projects
Multiple Speaker Speech Separation and Noise Separation using Conv-TasNet January 2024 - February 2024
Engineered and validated "Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation," utilizing Conv-TasNet for effective speech separation. Employed a deep learning model with a fully convolutional architecture for time-domain speech separation.
Enhanced computational speed, achieving 5 times faster processing than LSTM-TasNet, thus enabling real-time speech processing capabilities.
Improved performance by 5.8% in MOS subjective quality and 23.5% in PESQ objective quality compared to the LSTM model.
Increased efficiency, achieving significantly faster processing speeds, 5 times faster than LSTM-TasNet, facilitating real-time speech processing.
Boosted performance, realizing significant improvements in Source-to-Distortion Ratio improvement (SDRi) compared to previous STFT-based systems.
Using GANs to convert black and white images to color images January 2024 - February 2024
Pioneered a novel approach to colorize black and white movies using deep learning techniques, leveraging L1 loss and GAN loss to convert the task into a regression problem and enabling unsupervised machine learning solutions.
Attained the lowest Frechet Inception Score (FID) of 3.62 compared to other models, indicating more accurate color image generation.
Increased user preference by 46.3% compared to the previous model, producing more natural and vivid results.
Instant Violation And Control of Traffic Rule using Image Processing April 2022 - May 2022
Led a project to enhance road safety by automating helmet detection for bikers using machine learning, processing images captured through deployed cameras to identify and record violators.
Achieved a 98.4% accuracy rate in capturing images of violators, with a 94.0% F1 score.
Reduced processing time, achieving real-time processing capabilities compared to the previous model.
Ensured real-time performance while maintaining high accuracy, with a first-stage model size of 20.6 MB and a second-stage model size of 14.8 MB.
Demonstrated potential to reduce motorcycle accident fatalities by 37% by enforcing helmet use through this model.
Significantly reduced operational costs compared to traditional methods of helmet detection enforcement.
Experience
Graduate Assistant April 2023- May 2024
Ablah Library, Wichita, KS.
Collected and analyzed data to optimize the library's patron usage of computer resources.
Implemented improvements to streamline and optimize computer resource allocation, ensuring optimal functionality of all library computer systems, and increased computer usage by 30% using machine learning techniques for extracting insights from data.
Utilized A/B testing frameworks to optimize web content and user experience, and employed Google Analytics and Tag Management to track and analyze website performance.
Developed a dashboard to visualize key performance indicators such as patron count and PC usage, enhancing the library experience for patrons.
Internship at Eduvance December 2019- February 2020
Hyderabad, India
Developed standalone systems to monitor weather data and enable remote data retrieval.
Designed IoT-based systems to regulate temperature and moisture in soil within a greenhouse.
Conducted data analysis on collected data to make predictions and provide suggestions for system improvements.
Certificates And Awards
Held 2nd position in 36 hours Hackathon on Efficient use of green energy
Held 3rd position in 36 hours workshop on SMART-CITY HACKATHON
Complete A.I. & Machine Learning, Data Science Bootcamp by Udemy
Machine Learning Specialization by DeepLearning.AI
Deep Learning Specialization by DeepLearning.AI and Standford University
Deep Learning with Pytorch: Zero to GANs by Jovian