Nikhil Divi
Streetsboro,OH
Summary
An AI/ML Engineer with over one year of experience building and deploying machine learning solutions in healthcare, finance, and customer service. Developed a deep learning pipeline for melanoma detection using CNNs and transfer learning. Built NLP-based chatbots using transformer models like BERT, and implemented real-time fraud detection systems. Skilled in Python, PyTorch, TensorFlow, OpenCV, and cloud platforms (AWS, GCP). Strong background in MLOps, model optimization, and end-to-end ML lifecycle management.
Developed skills in data analysis and algorithm optimization within collaborative, fast-paced tech environment. Effective in translating complex data into strategic insights and solutions. Eager to transition to new field, bringing expertise in machine learning and passion for continuous learning and adaptation.
Overview
2
2
year of professional experience
Work History
Machine Learning Engineer
Beyond Soft Solutions
Hyderabad, India
12.2022 - 05.2024
- Developed a deep learning pipeline for melanoma detection, utilizing U-Net for lesion segmentation and Convolutional Neural Networks (CNNs) for classification across 7 types of skin lesions using dermoscopic images.
- Implemented automated disease classification achieving high sensitivity and specificity, with model performance competitive to professional dermatologists.
- Applied transfer learning with pre-trained CNN architectures (InceptionV3, ResNet, VGG19) to overcome data scarcity and reduce overfitting, improving model generalization across classes.
- Built multi-stage training pipelines, including data augmentation, image preprocessing, and class balancing to enhance prediction accuracy for rare skin cancer types.
- Integrated softmax-based output layers and cross-entropy loss functions to support multi-class classification; employed model fine-tuning and hyperparameter optimization for performance tuning.
- Achieved over 85% accuracy in classifying malignant vs. benign lesions, with metrics validated on a test set partitioned from ISIC datasets.
- Conducted exploratory data analysis (EDA) and feature engineering for dermoscopic features, enhancing interpretability of CNN model decisions.
- Deployed trained model components in a modular framework, facilitating future enhancements for real-time inference, web-based diagnostic tools, or integration into teledermatology platforms.
- Contributed to early detection solutions in digital health, highlighting the use of AI to reduce melanoma fatality rates through timely, cost-effective diagnostics.
- Developed and tested machine learning models to improve prediction accuracy.
- Collaborated with cross-functional teams to gather data requirements for model development.
Education
Master of Science - Computer Science
Kent State University
Kent, OH05-2026
Bachelor of Technology - Computer Science
Vel Tech University
Chennai, India05-2023
Skills
- Experienced in data analysis using Python, R, Scala, and SQL
- Machine learning frameworks
- Visualization: Matplotlib, Seaborn, Tableau, Power BI
- Experience with BERT architecture
- Libraries/Tools: spaCy, NLTK, Transformers (Hugging Face), Spark, Kafka, Flask
- Additional: Git, Jira
- Text analysis expertise
- Feature engineering
Languages
English
Full Professional
Timeline
Machine Learning Engineer
Beyond Soft Solutions
12.2022 - 05.2024
Master of Science - Computer Science
Kent State University
Bachelor of Technology - Computer Science
Vel Tech University