Dawood Sarfraz

FAST National University of Computer and Emerging Sciences

Bachelor of Science in Computer Science;

September 2020 - September 2024 Pakistan

Field of Specialization: Artificial Intelligence

Relevant Coursework:

• Probability and Statistics
• Calculus amd Analytical Geometry
• Differential Equations
• Linear Algebra
• Numerical Computing
• Artificial Intelligence
• Natural Language Processing
• Deep Learning
• Applied Machine Learning
• Data Science
• Digital Image Processing

Punjab Colleges

FSc Pre-Engineering

Gujranwala, Pakistan

Relevant Coursework:

• Physic I & II
• Mathematics I & II
• Chemistry I & II

Deutics Global

Associate Machine Learning Engineer

Jan 2025 - Up-to-date, Full-time Lahore, Punjab, Pakistan (Onsite)

• Working on Live Streaming
• Video Survillance Systems

FAST NUCES

Machine Learning Engineer (Research Assistant)

September 2023 - September 2024, Full-time Pakistan (Onsite)

• Project: Developed a multi-class skin cancer classification system using the HAM10000 dataset, which contains a diverse collection of skin lesion images. The goal was to classify different types of skin cancer accurately.
• NasNet:
  • Achieved highest classification accuracy of 93% by leveraging the depth of the architecture. Fine-tuned the model to adapt to the nuances of the skin lesion dataset, ensuring precise classification results.
  • Precision: 94%, Recall: 93%, F1 Score: 93%.
• ShuffleNet:
  • Enhanced computational efficiency by using grouped convolutions and channel shuffling, which reduced the model size and improved processing speed. Despite the smaller model size, accuracy was maintained at 87%, making it ideal for resource-constrained environments.
  • Precision: 87%, Recall: 87%, F1 Score: 87%.
• CNN:
  • Achieved accuracy of 92% using a custom CNN model.
  • Precision: 92%, Recall: 92%, F1 Score: 92%.
• Data Preprocessing & Augmentation:
  • Developed and optimized data pipelines to handle variations in the dataset, increasing the robustness of the models and reducing overfitting.
• Outcome:
  • Gained hands-on experience in deep learning model development and optimization. Enhanced my skills in solving real-world problems, particularly in the healthcare domain with applications in image classification for medical purposes.

Anonymous Tree

Machine Learning Engineer

Jun 2023 – Aug 2023, Internship Remote, Pakistan

• Project: Recommendation System Development
• Objective: Built a personalized recommendation system leveraging user data, item metadata, and advanced machine learning techniques to offer tailored product suggestions.
• Key Responsibilities and Tasks:
• Data Collection & Preprocessing:
  • Collected and processed user interaction data (e.g., clicks, purchases, ratings) and item metadata (e.g., product categories, descriptions, attributes).
  • Cleaned, normalized, and structured the data for efficient use in model training.
  • Implemented techniques like data imputation and feature scaling to handle missing values and ensure accurate model predictions.
• Model Development:
  • Developed a content-based recommendation system, utilizing item metadata (e.g., text-based features) to recommend similar products to users based on their previous interactions.
  • Implemented collaborative filtering techniques, such as user-based and item-based approaches, to suggest products based on user behavior patterns and item similarities.
  • Combined both content-based and collaborative filtering to form a hybrid recommendation system, improving the robustness and accuracy of the suggestions.
• Model Optimization:
  • Tuned hyperparameters to improve model performance using techniques like Grid Search and Randomized Search.
  • Implemented evaluation metrics such as Precision, Recall, F1 Score, and Mean Squared Error (MSE) to measure the effectiveness of the recommendation system.
  • Applied cross-validation to ensure the model’s generalization across different user segments and avoid overfitting.
• Algorithm Implementation:
  • Integrated Matrix Factorization methods (e.g., Singular Value Decomposition (SVD), KNN ) for collaborative filtering to predict missing interactions between users and items.
  • Utilized advanced techniques like TF-IDF (Term Frequency-Inverse Document Frequency) and Cosine Similarity for the content-based filtering approach to identify similar items based on their features.
• Performance Monitoring & Continuous Improvement:
  • Monitored the recommendation system's performance in real-time to ensure it met the business requirements and user expectations.
  • Collected user feedback and incorporated it into model retraining and improvements.
  • Analyzed system performance over time, identifying areas for enhancement and testing new features and algorithms to increase recommendation accuracy.
• Outcome:
  • Contributed to the development of a personalized recommendation engine that significantly improved product discovery and user engagement.
  • Enhanced the user experience by suggesting relevant products based on individual preferences, which led to increased user satisfaction and sales.

Natural Language Processing

Our NLP expertise gives your business next-level digital assistance that is contextually relevant, understands the language of people they use to communicate, and makes better decisions. We extract meaning from raw audio and text data to advance your NLP project.

Text Classification

We extract machine-readable information from unstructured text to enable data-driven approaches towards managing content and gain valuable insights from them.

Information Extraction

Document Processing

Sentiment Analysis:

With our cutting-edge machine learning algorithms, we uncover hidden meanings in vast amounts of text, effortlessly detecting your customers' sentiments. This valuable insight helps to enhance your business performance.

Social Media Oversight

Brand Surveilance

Market Analysis

strategize Workflow Enhancements

Machine Translation:

Our Machine Translation process automates the system, translating text or converting audio to text from one language to multiple languages without human intervention.

Automated Text Translation

Speech to Text

Text to Speech

Text Generation

Automated Transcription

Automated Chat-bots

House Prices - Advanced Regression Techniques

Month Year - Month Year

Contributor

• Description of your contributions and involvement in the project.

Languages

Technologies

Libraries