Esha Mukhtar

Results-driven professional with expertise in predictive modeling, data visualization, and end-to-end data pipeline development. Proven ability to transform raw data into actionable insights through machine learning and interactive dashboards. Passionate about applying AI to solve real-world problems and improve decision-making.

AI Developer at Forward Sports Private limited (Present)

• Automated Football Defect Detection: Developed an AI system to automatically recognize and classify football defects such as material issues, printing errors, and cutting defects.
• Utilized computer vision techniques for real-time defect detection and sorting, improving quality control in the manufacturing process.
• AI Boutique designs: Generated new boutique designs using Generative Adversarial Networks (GANs). Utilized Python (Pandas, Scikit-learn) and Flask for development. Applied deep learning techniques for creating unique designs and Focused on enhancing the user experience by providing AI-generated fashion items

Data Scientist at CodXo (2024-07 – 2024-12)

• SpaceX Data Collection and Wrangling: Scraped launch records using the SpaceX REST API and Wikipedia via BeautifulSoup. Transformed raw data into structured DataFrames for downstream analytics and dashboard use.
• SQL-Based EDA with Jupyter Notebooks: Performed in-depth SQL exploratory data analysis inside Jupyter Notebook. Used joins, aggregations
• Web Scraping for Product Data: Created web scraping scripts to extract structured product data from online marketplaces. Parsed tables and product stats into CSV format for later analysis and visualization.

Data Engineer | Data Analyst at Freelance work (2023-02 – Present)

A well-reputed private university in Pakistan offering specialized degrees in data science and AI

• Built ML models (SVM, k-NN, Naive Bayes) for fault detection in induction motors, achieving 92% prediction accuracy on sensor data.
• Ensemble learning techniques for medical image classification, improving diagnostic performance by 15% over baseline CNN.
• Real-time COVID-19 CT scan classification model using CNN, achieving 95% accuracy; contributed to public health analytics.
• Published findings in final-year research documentation; actively participated in faculty-supervised research reviews and poster presentations.
• Covid-19 Dashboard Analysis (India): Developed interactive Tableau dashboard showing vaccine doses by age, gender, and state. Informed public health strategy and outreach.
• Business KPI Dashboard (Power BI): Created performance dashboard tracking orders, shipments, and product KPIs.

Computer Vision /AI ML Engineer at Freelance / Personal Projects | GitHub (2021 – Present)

• Unsupervised Domain Adaptation (UDA) using Python and PyTorch for implementation, leveraging pretrained CLIP and large-scale image-text datasets.
• Unsupervised Domain Adaptation (UDA) System using CLIP and modality separation, enabling cross-domain image classification and recognition without labeled target data.
• Canny Edge Detection & Image Processing (Custom Implementation): Applied Gaussian blur, gradient calculation, non-max suppression & hysteresis.
• Facial Recognition using HOG & Template Matching: Implemented a custom facial recognition system using Histogram of Oriented Gradients (HOG) and sliding window matching.

Bachelors in Data Science – GIFT University (2020 – 2024)

Generative AI – Coursera (2024-12 – 2025-06)

Talend Data Integration – Coursera (2024-12 – 2025-06)

Certified Data Analyst – Simplilearn (2024-10)

Python Developer – Sololearn (2022-11)