Ishan Islam

Long Island University, Brooklyn, NY

B.S. in Computer Science, Expected May 2026

• GPA: 3.68/4.00
• Vice President of CS Club
• Vice President of Google Developer Group LIU
• Dean's List Student

Accenture | AI and Machine Learning Fellow

Aug 2025 – Present

• Contributing to a fake news classification project, developing NLP pipelines and deep learning models to detect misinformation across large-scale datasets.
• Preprocessing and engineering features from unstructured text data, improving model performance and reducing noise in classification tasks.
• Implementing and fine-tuning supervised learning models using Scikit-learn, TensorFlow, and PyTorch, achieving measurable improvements in accuracy and precision.
• Designing data visualization dashboards with Python and Tableau to communicate model insights and performance metrics to both technical and non-technical stakeholders.
• Assisting in optimizing end-to-end ML workflows, including data cleaning, model training, hyperparameter tuning, and deployment in test environments.

Break Through Tech | AI and Machine Learning Fellow

Jun 2025 - Present

• Engaging in an intensive Machine Learning Foundations Curriculum, working with industry-relevant tools to analyze real-world datasets and develop data science projects.
• Participating in AI Studio and Portfolio Development, solving real-world ML challenge projects under the guidance of academic and industry experts.
• Receiving mentorship and career coaching, collaborating with industry professionals on case studies, skill-building simulations, and personalized feedback sessions.

Long Island University | Techincal Assistant

September 2023 – Current

• Diagnosed and resolved technical issues on computers to ensure optimal performance.
• Coordinated with service providers to schedule equipment maintenance and repairs.
• Performed software updates on library computers and confirmed proper hardware functionality.

Long Island University | Google Developer Group Vice President

Sept 2024 - Present

• Streamlined student onboarding by automating email invitations through SMTP.
• Secured hundreds of dollars in scholarships from BWOSS for club members, enabling them to obtain Google Career Certificates at no cost
• Partnered with Major League Hacking to host successful hackathons with over 20 attendees.

TikTok Video Classification

Lead Data Scientist

• Eliminated 100% of inaccuracies in media details through data cleaning with the Python data stack (Pandas and Numpy)
• Developed a classification model, using to predict whether TikTok video transcriptions were claims or opinions, based on text features
• Tuned hyperparameters for Random Forest and XGBoost classifiers using GridSearchCV, optimizing for recall
• Preprocessed text data by extracting numerical features using CountVectorizer.
• Evaluated model performance through classification reports, confusion matrices, and feature importance plots, achieving strong recall and precision.
• Evaluated model performance through classification reports, confusion matrices, and feature importance plots, achieving strong recall and precision.
• The final model was deployed on test data, where the Random Forest model provided key insights into feature importance.

Waze Regression Analysis

Lead Data Scientist

• Developed a logistic regression model to predict user churn based on driving behavior and app usage metrics
• Cleaned and processed the dataset by handling missing values, scaling features, and capping outliers.
• Through exploratory data analysis (EDA), including correlation heatmaps and descriptive statistics, I gained insights into the data.
• Built and optimized a logistic regression model, evaluating its performance with metrics like precision, recall, F1-score, and a confusion matrix
• Created visualizations to determine results

Lightning Strike Analysis and Geological Graphing

Lead Data Scientist

• Developed an interactive Tableau dashboard to analyze the frequency, intensity, and patterns of lightning strikes across various geographical regions
• Integrated geological maps to overlay environmental features such as elevation, vegetation, and proximity to water bodies, correlating them with lightning activity.
• Utilized advanced Tableau functionalities, including calculated fields, filters, and parameters, to enable real-time data exploration and hypothesis testing.
• Presented findings on lightning-prone areas, offering actionable insights for disaster preparedness and risk management strategies.