I've built ML powered CRM systems that increased conversion rates by 20% and multimodal medical AI achieving near-radiologist performance.
My expertise includes cybersecurity voice cloning systems for penetration testing and Azure-based sentiment analysis pipelines with Event Hubs and Databricks.
My background spans business, psychology, and data science, with experience bridging technical solutions and business impact.
PyTorch, scikit-learn, OpenCV, Hugging Face, ResNet, VGG16
Computer vision is revolutionizing industries from autonomous vehicles to manufacturing, with healthcare diagnostics seeing particularly transformative impact as AI systems detect diseases from medical imaging at scale.
My chest X-ray diagnostic system used fine-tuned vision models to achieve near-radiologist accuracy, generating clinical summaries across 220K+ images
Python, Hugging Face, Real-time STT/TTS, LLM APIs, Multi-agent orchestration
LLM-powered agents are creating sophisticated new attack vectors in cybersecurity, with AI-generated social engineering becoming one of the most difficult threats to defend against.
I've designed a security penetration testing system that orchestrates autonomous phone calls by coordinating voice cloning, real-time speech transcription, and LLM-driven conversational agents that respond dynamically to targets
Azure Event Hubs, Databricks, Azure Data Lake, Python, PySpark, LLM APIs
Streaming data architectures enable organizations to respond to emerging trends as they happen rather than discovering them days later in batch reports.
An Azure pipeline I developed processes live social media streams for LLM-based sentiment analysis with automated trend reporting
Python, SQL, REST APIs, Docker, Git, Azure, Cross-domain Architecture
The best solutions emerge from experimenting across domains and connecting technologies in novel ways.
My work spans healthcare, cybersecurity, and analytics - each requiring creative integration of technologies to solve cross-domain challenges.
Azure-based data pipeline with real-time LLM sentiment analysis for trend reporting. Built with Event Hubs and Databricks for scalable processing.
Automated chest X-ray classification using ResNet, VGG16, and LLaMA vision models. Achieved near-radiologist performance on diagnostic accuracy.
Replacing traditional voice assistants with a custom GPT pipeline on Raspberry Pi, integrating LLM-powered natural language control with smart home devices for superior conversational experiences.
Building a fully autonomous conversational system for penetration testing that generates real-time voice cloning attacks to audit social engineering vulnerabilities in security protocols.
Developing a browser extension that uses machine learning to identify and filter spam content in real-time, protecting users from phishing attempts and malicious communications.
I see problems everywhere - not in a pessimistic way, but as puzzles waiting for solutions. Take home automation: the entire industry has ignored LLMs even though they provide vastly superior experiences to preprogrammed voice commands. So I'm building my own - replacing Alexa with a Raspberry Pi running a custom GPT pipeline I can SSH into and integrate with my house. This isn't a school project or portfolio piece; I'm building it because I want to actually use it, and the technical challenge itself is what drives me.
My background spans business, psychology, and data science, which means I naturally think about problems from multiple angles: user behavior, business metrics, and technical feasibility. I've built ML-powered CRM systems that increased conversions by 20%, medical imaging classifiers approaching radiologist accuracy, and cybersecurity voice cloning systems for penetration testing. Each project pushed me into unfamiliar territory - healthcare regulations, real-time audio processing, cloud architecture - and that's exactly what made them interesting.
I'm currently finishing my Master's in Data Science at Northeastern University, but my real education happens in side projects: home automation pipelines, multimodal translation systems, whatever problem catches my attention. I'm looking for opportunities where building genuinely useful solutions matters more than checking boxes on a tech stack wishlist.
Ready to discuss your next project? I'd love to hear from you.