Skip to content

Jayasimha Obulam

Data Scientist • Kadapa, Andhra Pradesh, India • o**************@gmail.com • +91*******539 • drivetube.ai/•••••

Professional Summary

Data Scientist with 0 years of experience building and fine-tuning machine learning models using Python and deep learning frameworks. Applied transfer learning and convolutional neural networks to medical imaging, achieving 95% accuracy on a 12.5k-image Kaggle blood cell dataset. Skilled in TensorFlow/Keras, data preprocessing, augmentation, hyperparameter tuning, and model evaluation; seeking entry-level roles to apply ML and computer vision skills to healthcare or applied AI problems.

Technical Skills

Programming Languages: Python
Frameworks and Libraries: TensorFlow,Keras,Scikit-learn,NumPy,Pandas,Matplotlib,Seaborn
Databases: SQL
Cloud and DevOps: AWS
Data and Analytics: Transfer Learning,Convolutional Neural Networks,Model Fine-Tuning,Deep Learning,Supervised Learning,Jupyter
Tools and Methodologies: Git,Kaggle,GitHub
Concepts & Methods: Data Preprocessing,Data Augmentation,Hyperparameter Tuning,Model Evaluation

Work Experience

SmartBridge
Data Science Intern
May 2025 – Aug 2025
Worked on medical imaging ML for automated blood cell classification using transfer learning and CNNs to support faster, more accurate microscopic analysis.
Tech Stack: Python, TensorFlow, Keras, NumPy, Pandas, Scikit-learn, Jupyter, Git, Kaggle
  • Fine-tuned pre-trained convolutional neural network architectures using TensorFlow and Keras on a 12,500-image Kaggle blood cell dataset, achieving 95% classification accuracy and outperforming a baseline CNN trained from scratch.
  • Implemented an image preprocessing and augmentation pipeline (normalization, rotations, flips, scaling) in Python and NumPy to increase dataset diversity and improve model generalization.
  • Performed hyperparameter tuning (learning rate schedules, batch size, early stopping) and training optimizations to accelerate convergence and reduce overall training time across experiments.
  • Conducted model evaluation and error analysis using confusion matrices and per-class precision/recall to identify under-performing classes and guide targeted augmentation strategies.
  • Packaged the trained inference pipeline in reproducible Jupyter notebooks and version-controlled code (Git), delivering a faster, more accurate alternative to manual microscopic review and reducing analyst workload.
  • Documented experiments, tracked model artifacts and results to enable reproducibility and handover to cross-functional stakeholders for potential integration into clinical workflows.

Projects

HematoVision: Advanced Blood Cell Classification Using Transfer Learning | 2025 – 2025
Tools Used: Python, TensorFlow, Keras, NumPy, Pandas, Data Augmentation
  • Designed and implemented an end-to-end image classification pipeline using transfer learning with pre-trained CNNs (TensorFlow/Keras) to classify microscopic blood cell images from blood smear data.
  • Achieved higher classification accuracy and faster training convergence compared to a baseline CNN trained from scratch by applying transfer learning, preprocessing, and augmentation techniques and validating on a held-out set.
  • Prepared reusable code, model checkpoints, and inference notebooks to demonstrate scalability and support potential integration into healthcare analysis workflows.

Education

Kandula obul reddy memorial college of engineering
Bachelor's Degree, Computer Science / Engineering • Oct 2022 – Apr 2026
Coursework: Data Structures & Algorithms, Machine Learning, Statistics, Database Management Systems (SQL), Python Programming

Powered by Drivetube · Create your own profile at drivetube.ai