Imran Khan

● Led a project to develop predictive models for customer churn using deep learning frameworks, achieving 95% prediction accuracy. ● Deployed scalable machine learning solutions on Azure, leveraging CI/CD pipelines for automated updates. ● Worked extensively on data preprocessing and visualization to prepare and present actionable insights to stakeholders. ● Designed ETL pipelines for large-scale data management, streamlining data preparation and storage processes.

As a freelance graphic designer, I offered creative design services tailored to clients’ needs, focusing on logo design, branding, print media, and digital content creation. I also specialized in children's storybook design, outsourcing the artistic aspects to skilled illustrators and ensuring high-quality final products. Logo Design: Crafted unique logos for businesses in diverse sectors, ensuring each logo captured the brand’s essence. Branding: Developed complete branding packages with consistent logos, color schemes, and typography for businesses seeking cohesive visual identities. Business Cards & Stationery: Designed professional business cards and stationery, aligning with clients’ brand aesthetics. Social Media Graphics: Created custom graphics for social media platforms, supporting clients’ marketing campaigns with engaging visuals. Marketing Materials: Designed brochures, flyers, and posters, enhancing brand awareness and attracting new customers. Children’s Storybook Design: Worked on children’s storybook projects where I collaborated with skilled artists for illustrations, managing the entire design process, from layout to final artwork. I handled the book layout, typography, and design direction, ensuring that the illustrations and text were integrated seamlessly. Outsourcing: For highly artistic tasks such as illustration, I outsourced to skilled freelance illustrators, ensuring that the work met the highest standards while maintaining project timelines and budgets. I pride myself on delivering creative and high-quality work that exceeds expectations.

● Built and deployed machine learning models for customer behavior prediction, increasing marketing campaign efficiency by 20%. ● Designed interactive dashboards in Tableau to visualize sales and operational data, enabling data-driven decision-making across departments. ● Developed and managed data pipelines using Python and SQL, ensuring seamless integration with cloud platforms (AWS). ● Conducted model optimization and fine-tuning, improving accuracy and reducing runtime by 25%.

Machine Learning project at Computer Science Department

AI Certification with A+ plus grade - Topper of the entire batch

➤ Modules: Data Science, Big Data Analytics, Leadership, Information Governance and Cyber Security, Research Methods and Machine Learning on Cloud. ➤ Master's Project Led an ASR project to improve voice recognition accuracy using deep learning techniques. Worked with the LJSpeech dataset to optimize model performance. Key Responsibilities: ➤ Data Preprocessing & Feature Engineering • Preprocessed audio data and applied feature extraction (MFCC, spectrograms). • Implemented data augmentation to improve model robustness. ➤ Model Development & Optimization • Built and optimized models (RNN, CNN, Transformer) to reduce Word Error Rate (WER) by 5-6%. • Tuned hyperparameters to enhance model efficiency. ➤ Performance Evaluation & Problem-Solving • Evaluated model using WER and Signal-to-Noise Ratio (SNR) metrics. • Improved model accuracy by 8-10% in noisy environments. ➤ Results • Reduced WER by 6% and improved recognition in noisy settings by 8%.

Bachelor’s Project: Personality Prediction System Using Text and Image Processing: Developed a desktop application to predict personality traits by analyzing text and image data using machine learning and deep learning techniques. Key Features: Text-Based Prediction: Used Naive Bayes and SVM classifiers to predict traits like extraversion and openness from text. Face Detection & Emotion Recognition: Integrated OpenCV for face detection and VGG-16 for recognizing emotions in faces, linking them to personality traits. Multimodal Analysis: Combined text and image data to enhance the accuracy of personality predictions. Evaluation & Results: Accuracy: Achieved 85% overall prediction accuracy. Text Analysis: 80% accuracy for classifying traits. Emotion Recognition: 85% accuracy for detecting facial emotions. Multimodal Integration: Improved accuracy by 5-10% when combining both text and facial data.