Principal Data Scientist

I led a project focused on quantifying risk in monetary terms for a chain of hospitals in the USA. Utilizing advanced techniques such as Probabilistic Determination, FAIR STRIDE, Monte Carlo Simulations, and Statistics, we developed a cutting-edge solution. The result was a product that provided real-time updates on risk and likelihood, taking into account Common Vulnerabilities and Exposures (CVEs) as well as internal risk factors.

I spearheaded a project that revolutionized query creation and visualization for databases. Leveraging state-of-the-art Language Models (LLMs), we achieved an impressive 100% accuracy in dynamically generating queries for Cassandra, MongoDB, and MYSQL DB. By eliminating canned queries and introducing flexible querying capabilities, we empowered users to extract any information they needed from the databases.

I focused on real-time threat detection for hospitals, utilizing IoT and network data. The project involved tackling a time series anomaly detection problem through deep learning techniques. By leveraging cutting-edge technologies, we successfully implemented a real time threat identification and mitigation product to ensure the security of hospital networks.

In collaboration with Yale University and ChestAi, I worked as an Advisor Data Scientist to develop a product capable of detecting chest diseases through X-ray images. My contributions involved enhancing the accuracy of the product by solving computer vision problems. Additionally, I played a role in product marketing, acquiring test images, and onboarding hospitals for beta testing. This project aimed to provide an efficient and reliable diagnostic tool for healthcare professionals.

As a Co-Founder and Data Scientist at DermaAI, I led the development of a computer vision solution for the accurate detection of malignant, chronic, and cosmetic skin conditions. By leveraging state-of-the-art computer vision techniques and analyzing various medical images, our product surpassed the accuracy of dermatologists. I was involved in all stages of the project, from idea inception to data collection, data science implementation, and sales.

As a Senior Machine Learning Engineer at Intel Telecommunication Deutschland and LNT Technology Services, I played a key role in classifying telecommunication logs. This involved tasks such as data ingestion, data wrangling, and designing a database schema. By developing an accurate model that achieved 97% classification accuracy, we successfully automated the process of directing relevant logs to the appropriate analysis teams. The implementation of our data science solution resulted in significant time savings of 3,200 person-hours per month for the client.