Developing AI agents & Jotform Copilot (AI Form Builder).- RAG Pipeline Implementation- Model Fine Tuning - Structured Output Generation- Generating Advanced Reasoning Models via RL, CoT and ToT

Data Scientist at Fraud & Abuse AI team

Intern at Radar, Electronic Warfare & Intelligence Systems.Worked on a web based application as a full-stack developer.

Created a Recommendation System for Form Templates: Using language tools and methods such as NLP (Natural Language Processing) and NLTK, I developed a system that suggests form templates. PyTorch was used for building the model, and I carefully cleaned the data to make it suitable for deep learning. This involved using sentence transformers to make sense of complex language. Built a Multilingual AI Chatbot: I designed an AI chatbot that can understand and communicate in 25 different languages. By using a Naïve-Bayes classifier and techniques to recognize names and other entities, the chatbot can accurately figure out what users are asking for based on their prompts. Developed an Interactive Frontend for the Recommendation System: Using React.js, I created a frontend that responds to user needs and provides a lively customer experience for the form template recommendation system. Constructed a High-Performance Backend: I used FastAPI to develop the backend of the system, ensuring smooth and efficient interactions with the data. This provides a high level of performance for the entire application.

Tutor of CS 101 - Algorithms and Programming I course

Tutor of CS 125 - Introduction to Data Analytics for Social Sciences course

Tutor of CS 102-Algorithms and Programming II course

-Taken role in a project which tries to determine payback ratio of the customers who are in debt over a period of time. - Taken role in a project to determine the relation between digital banking usage and credit card customers between 2019-2022.