At Careem, as a Software Engineer II in MLOps, I spearhead the development and deployment of an innovative in-house translation solution utilizing Large Language Models (LLMs). This solution has not only enhanced translation accuracy by 10% but also increased output efficiency by 97% over the existing systems, while localizing content from English to Arabic, significantly improving grammatical accuracy and preserving the integrity of intended meanings.In my role, I extend my expertise to optimizing the accuracy and scalability of our translation services. I lead a rigorous quality assurance process that incorporates continuous human feedback to ensure and enhance translation accuracy across various domains. This direct interaction with stakeholders ensures that our solutions effectively meet diverse user needs.A key achievement during my tenure at Careem has been the localization of over a million strings across various business domains, demonstrating my commitment to delivering robust localized experiences at scale. This milestone not only highlights the technical capabilities of our solutions but also underscores their practical impact on the market.Throughout this journey, I have honed valuable skills in Python, FastAPI, AWS, and microservice architecture, alongside mastering tools like LLM and Langchain. My interactions with various teams and stakeholders have allowed me to adapt and refine our solutions and architecture to better meet their needs. My career at Careem has been characterized by continuous growth, driven by relentless innovation and a commitment to embracing cutting-edge technologies. This experience has refined my skills and fueled my passion for developing impactful, real-world solutions.

At Confiz Limited, as a Senior Machine Learning Engineer, I led a dynamic team of machine learning engineers in drastically reducing model pipeline execution times—from 5 days to under 12 hours. This achievement significantly accelerated project delivery and showcased our team's capability in enhancing operational efficiency.Our collaborative efforts extended to working closely with the Data Science team, where I oversaw the integration of machine learning perspectives into their analyses. We designed and implemented a reproducible ML pipeline, which included a robust architecture for tracking artifacts and ensuring scalability. This initiative was critical in maintaining high standards of data integrity and reliability.We implemented stringent data profiling and quality checks at every step of the pipeline, which reduced error rates by 40%. My team was also responsible for developing and managing ETL jobs that streamlined the delivery of forecast data to the business team, thereby optimizing the decision-making process. A key achievement in our tenure was the development of a pipeline that automatically generates and delivers accurate forecast results every month to stakeholders, enhancing strategic business planning.Throughout these initiatives, we leveraged advanced tools and technologies including Python, PySpark, Google Cloud, and Dash/Plotly. These tools enabled us to develop interactive dashboards and perform complex data processing tasks efficiently, further improving our markdown forecasts and operational capabilities.Through these efforts, my leadership in these projects not only advanced the goals of the ML Competency at Confiz Limited but also underscored our commitment to delivering robust, scalable machine learning solutions.

At Confiz Limited, as a Machine Learning Engineer, I maintained and fine-tuned a customized PySpark pipeline to train various models for generating forecasts. I employed industry-leading methodologies to automate the ML pipeline execution and ensure seamless integration, while also enforcing continuous integration for the ML pipeline codebase. By leveraging best ML practices, I structured the pipeline results stored in BigQuery into customized models, enabling on-demand forecasting.My role also involved designing and implementing bespoke model registry modules through API calls, which enhanced model binning and registration, thus accelerating model deployment and improving version control.My contributions extended to designing and executing multiple ETL jobs that shared pipeline results with the business intelligence team. These efforts significantly improved the runtime of multiple ETL batch jobs processing millions of records by 95%, thereby supporting efficient business intelligence operations.A notable achievement during my tenure was the orchestration of the ML pipeline to run in an automated way, showcasing the pipeline's capability for large-scale efficient data processing. I also played a key role in the structuring of the Machine Learning Competency and the training of ML resources, ensuring the team's continued growth and skill development.Throughout this period, I honed valuable skills in Python, PySpark, Google Cloud, BigQuery, and CI/CD processes. These technical proficiencies enabled me to contribute significantly to the team's success and the overall efficiency of our machine learning operations.

At Confiz Limited, as a Software Engineer, I developed efficient REST APIs that served over a million clients daily, achieving an impressive average response time of less than 300 milliseconds. My work significantly improved customer satisfaction by optimizing the performance and reliability of these services.I played a key role in transitioning legacy systems to a scalable microservices architecture using state-of-the-art technologies such as Spring Boot, Kafka, Docker, and Azure. This transition enhanced the maintainability and scalability of our systems, ensuring they could handle increasing loads and evolving requirements.To ensure our services were production-ready, I implemented continuous integration and continuous deployment (CI/CD) processes for the APIs. Additionally, I wrote integration tests, incorporated comprehensive logging, and integrated monitoring and alerting dashboards using Splunk. These efforts reduced downtime by 30% and enhanced system reliability and performance.I was also involved in planning, proposing, and delivering proofs of concept for new features. My contributions included conducting detailed feasibility studies and leading the implementation of innovative functionalities, which enabled successful new feature deployments.Throughout this period, I honed valuable skills in Spring Boot, web services, Docker, and CI/CD processes. My technical expertise and proactive approach allowed me to contribute significantly to the team's success and the overall improvement of our software engineering practices.

As an Ambassador for deeplearning.ai, I engaged the deep learning community in Pakistan by organizing AI-related events in Lahore. My efforts fostered a collaborative environment, promoting the growth and knowledge-sharing within the AI community.

As a Web Development Teacher Assistant at the University of the Punjab, I supported the web development course by creating quizzes, preparing and evaluating lab exercises, and conducting vivas. My role contributed to enhancing the learning experience for students and ensuring the smooth delivery of the course content.

As a freelancer on Fiverr, I developed a full-stack web application using JSP for a Bahrain-based company. This project included creating an admin panel for website management, inventory creation, and report generation. Additionally, I developed an image template editor using FabricJS to dynamically fill template content, enhancing the user experience and functionality of the web application.