cv

Software Engineer Intern

Social Media Lab, Toronto Metropolitan University

Contributed to the PoliDashboard server by integrating LLM-powered features and conducting data analysis for research papers.

• Integrated summarization features using Hugging Face Transformers, Ollama, and OpenAI APIs.
• Benchmarked multiple LLMs for performance, accuracy, and latency.
• Authored literature reviews and technical reports supporting peer-reviewed research.

Computational Biology Research Assistant

Biomathematics and Fluids Group, Toronto Metropolitan University

Implemented predictive HIV models to simulate viral rebound after treatment, utilizing solver optimization to reduce simulation time.

• Developed models in Julia using DifferentialEquations.jl.
• Conducted Markov Chain Monte Carlo (MCMC) simulations for parameter estimation and uncertainty analysis.
• Presented research findings at Complex Systems Day at TMU.

Co-op Program Research Assistant

Career Co-op & Student Success Center, Toronto Metropolitan University

Analyzed recruitment data to generate hiring trend reports and optimized digital resources for student use.

• Analyzed over 250 job applications using Python (Pandas) and MySQL.
• Redesigned the co-op skills portal, improving page load times by 40%.
• Secured 5+ internship partnerships through coordination with industry and faculty.

Vice-President

Toronto Metropolitan Coding and Algorithms Club

Leading the development of technical programming resources for the student body.

• Developed and led technical workshops focused on data structures, algorithms, and mock interviews.

Bachelor of Science

Toronto Metropolitan University

Computer Science, Minor in Biology

From Regression to Classification: Exploring the Benefits of Distributional Representations of Energy in MLIPs

CVIS 2025 Conference Submission

This paper proposes training Machine Learning Interatomic Potentials (MLIPs) using a histogram-based cross-entropy loss instead of scalar regression[cite: 1]. [cite_start]By predicting energy distributions, the model achieves accuracy comparable to regression baselines while enabling the quantification of epistemic uncertainty through prediction entropy.

HIV Drug Discovery Data Integration

Collected and standardized bioactivity data for HIV-1 Integrase and RNaseH to identify structure-activity trends.

• Standardized molecular activity values for cross-dataset comparability.
• Engineered molecular fingerprints using RDKit to explore chemical space.
• Built Random Forest and XGBoost models to predict compound potency.

Emotionally Responsive Robot with Multimodal Feedback

Designed a robotic system capable of interpreting user inputs to generate dynamic emotional feedback.

• programmed an Arduino Mega to generate expressive RGB-LED patterns based on a rule-based personality model.
• Co-authored a report on applying affective computing to autonomous robotics.