AI & Machine Learning Researcher with 7+ years of experience architecting scalable, real-time AI and IoT systems. Specialized in Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and intelligent frameworks for real-time IoT data processing and search. Passionate about advancing inclusive, impactful research in AI and Generative Systems.
Contributed to the success of the University of Toronto's online Data Analytics Boot Camp specializing in training data professionals.
GPA: 4.22/4.3. Thesis: SensorsConnect: World Wide Web for Internet of Things.
Thesis: Multi-modulation Low Earth Orbit Satellite Based on Software Defined Radio.
Capstone: EagleEyes – Mine Detection System using a Quadcopter (Drone).
Leveraged the GPT-2 language model and the Hugging Face Transformers to fine-tune GPT-2 on news classification tasks using the AG News dataset. Implemented a parameter-efficient fine-tuning (PEFT) approach. Achieved improvements in the evaluation metrics, including reduced loss and increased classification accuracy.
View Project
Developed a real-time IoT search engine powered by LLMs and RAG, supporting natural language queries across heterogeneous IoT systems. Implemented a semantic search pipeline using Sentence-BERT and HNSW indexing. Managed over 37,000 real-time IoT documents across 500 service types stored in MongoDB with geographic indexing. Achieved 92% top-1 accuracy in complex intent detection and retrieval, outperforming state-of-the-art systems like Gemini. Applied in real-time urban scenarios: finding least-crowded clinics, available parking, and lowest gas prices based on live sensor data.
View Project
Developed RadViz-Plotly, an open-source Python package for creating 2D and 3D Radial Visualization (RadViz) plots for high-dimensional datasets. Enabled data scientists to visualize complex data distributions interactively using Plotly for enhanced insight discovery. Facilitated better understanding of high-dimensional data and detection of hidden patterns through intuitive visual analytics.
View Project