Hi there 👋
I'm Baljindersingh Surendrasingh Bedi!
Check out my portfolio.
🎓 Education:
- Master of Science in Artificial Intelligence from DePaul University, Chicago, IL (GPA: 3.5)
- CA Intermediate from The Institute of Chartered Accountants of India, India (GPA: 3.6)
- Bachelor of Commerce – Accounting, Finance, and Taxation from Gujarat University, India
👨💻 Professional Experience:
Artificial Intelligence Engineer Intern
RadicalX, New York | June 2023 – Aug 2023
- Advocated for open-source models to uphold data privacy standards, ensuring ethical AI development.
- Implemented FAISS for managing high-dimensional vector databases, improving data retrieval efficiency by 40%.
- Developed an AI system using the OpenAI API, enhancing information retrieval and user interaction.
- Improved question-answer formats, boosting accuracy and reducing errors by 40% and 25%.
- Conducted a comparative performance analysis between BERT and LLama2 models, showcasing LLama2's superiority.
🚀 Project Experience:
- Code Translator App: Developed a robust application capable of translating code between various programming languages (Python, Java, C++, etc.) as well as handling different databases (SQL, NoSQL). The app supports code conversion for multiple use cases such as function-based, object-oriented, and database-related code snippets. It utilizes large language models (LLMs) like Groq, OpenAI, and Llama3 for accurate translations. It also supports floating-point formats (FP16, FP32, bfloat16) and includes automatic handling of bash scripts, command-line utilities, and configuration files for seamless integration with different environments. Check out the GitHub repository.
- Hybrid 3D Pooling CNN for Gesture Recognition: Developed a CNN model for gesture recognition in low-resolution videos, achieving 93% accuracy.
- 3D Reconstruction using RGB and Depth Images: Developed a 3D reconstruction model as part of my thesis, focusing on the accuracy and efficiency of the reconstruction process.
- AI-Driven Dino Game Using CNN: Built an autonomous game player for the Chrome Dino game using CNN and reinforcement learning.
- Tutor Buddy: Created a personalized tutoring system using LLMs like LLama3 and Gemini, incorporating algorithms for adaptive learning.
- RAG Application for Chatbots: Designed a Retrieval-Augmented Generation system for chatbots, integrating FAISS for efficient data retrieval. Check out the GitHub repository.
- Bank Loan Default Prediction: Utilized various machine learning algorithms to predict the likelihood of loan defaults.
- Retail Rocket Recommender System: Analyzed e-commerce behavioral data and improved product recommendations using item-based, user-based, and hybrid models.
🚀 Skills & Tech Stack:
- Programming Languages: Python, Java, PHP, SQL
- AI & Machine Learning: TensorFlow, PyTorch, scikit-learn, Keras
- Frameworks & Libraries: Hadoop, Spark, Flink, FAISS, OpenCV, Tesseract OCR Engine
- Models & Techniques: Transformers, Large Language Models, Reinforcement Learning, Decision Trees, XGBoost
- Data Science & Analysis: Data Visualization, Feature Engineering, Data Preprocessing
- Version Control: Git
🌱 Continuous Learning:
- Exploring Rust and enhancing my knowledge in 3D CNNs and advanced ML techniques.
😄 Fun Fact:
- I love solving complex machine learning challenges and enjoy pushing the boundaries of AI applications.
🔍 Future Goals:
- To contribute to innovative AI projects and continue expanding my expertise in machine learning and AI-driven applications.
📫 Get in touch:
- Email: [email protected]