Machine Learning Engineer

Developed a Flask-based application dedicated to backtesting various trading strategies. This application
provided a simulation environment to evaluate the effectiveness of strategies using historical data, offering
insights on potential profits and risks. Through the application, users could fine-tune their strategies before
deployment, significantly enhancing decision-making and profitability in real-world trading.
 engineered a cross-platform mobile application for plant disease classification, utilizing Android Java for the app
development. The core of the application lies in its machine learning models, which leverage architectures like
VGG and ResNet along with the concept of transfer learning in convolutional neural networks (CNN). The
models were trained to recognize and classify various plant diseases effectively. The app and its machine
learning components were seamlessly deployed using AWS SageMaker, providing a user-friendly interface for
real-time plant disease diagnosis.
 Effectively integrated various technologies to optimize and deploy a highly efficient system. The core of the
system involved leveraging LLMS and Langchain for language modeling and analysis. I used Pinecone for
efficient vector search, and fine-tuned ChatGPT to cater to specific application requirements. The open-source
LLMS models were customized to fit our needs. All these components were orchestrated with AWS Lambda
functions, allowing the system to run serverless, which significantly improved scalability, performance, and
cost-effectiveness.
 In a significant project on deepfake technology, I utilized the Wave2Lip model, renowned for its performance in
generating lip-sync deepfakes. I fine-tuned this model on our specific dataset to improve its accuracy and
relevance to our unique use case. To ensure a smooth deployment and maintain version control, I encapsulated
the fine-tuned model into a Docker image. Finally, I deployed the Docker container using RunPod, a platform
known for its ease of use and scalability, enabling efficient application management and seamless integration
with our existing infrastructure.
 Constructed an extensive application focused on image classification and recognition by utilizing machine
learning algorithms. This application employed advanced neural network architectures such as VGG, ResNet,
and YOLO for efficient and accurate image processing. In this process, I used these models for feature
extraction, classification tasks, and object detection, generating significant insights from the data.
 Enhanced the training dataset through data augmentation using generative networks. This involves generating
new data that expands and diversifies the training set by creating transformed versions of existing data points.
 Orchestrated the entire process of a deep learning project from start to finish. This includes data collection and
preprocessing, model design and training, hyperparameter tuning, model evaluation, and final model
deployment As a Data Management Specialist for Bupa Arabia, I spearheaded the data management process using Azure
Purview, a unified data governance service.I bolstered data management by ensuring clear data lineage,
facilitating traceability and system transparency. I created a user-friendly data catalog, enabling efficient data
discovery and utilization, and developed a comprehensive data glossary to standardize understanding of data
elements. My focus on data governance best practices improved data accessibility, usability, and quality,
enhancing the overall value of our data assets.
 Developed a full-stack application for Monster for time-series data processing and analytics using Azure
services such as Synapse Analytics, Data Factory, and SQL Server Integration Services. Implemented machine
learning algorithms to generate insights from the data. Utilized Data Factory to orchestrate data pipelines and
Synapse Analytics to build and train machine learning models. Demonstrated proficiency in developing scalable
and robust solutions on Azure.
 Developed a computer vision project in which I leveraged my expertise in computer vision to develop a robust
C# console application. Using libraries such as Pillow and OpenCV, I designed and implemented various
algorithms to address image distortion and object length for TaylorMadeGolf Company. Specifically, I focused
on utilizing image warping techniques to retain CMYK image properties.I utilized my expertise in C#
programming and image processing to design and implement effective algorithmsOrchestrated the entire
process of a deep learning project from start to finish. This includes data collection and preprocessing, model
design and training, hyperparameter tuning, model evaluation, and final model deployment
 Backend Developer for Behaviorally web automation tool. Responsibilities included writing Power Automate
flows, scraping power bi reports using puppeteer, generating editable PPTs using Pptxgenjs deploying the tool
on Express server, using Azure Queue Storage and schedulers to cater to requests, assigning tasks, doing PR
reviews and POCs.