Machine Learning Team Leader
Resume:
BIBEK KUMAR SHARMA
*** * ***** **, *** A* Moscow, ID 83843 ***********.**@*****.*** https://www.linkedin.com/in/bibek-sharma-5b382a152/
University of Idaho Aug 2021 - Dec 2025
Bachelor of Science, Computer Science (GPA: 3.61)
•Coursework: Machine Learning, Software Engineering, Compiler Design, Data Structures and OOP, Machine Vision, AI Data Analysis, Database System, Data Science, Real-Time OS, Data Analysis, System Software, Computer Architecture
SKILLS
• Programming Languages: C/C++/C#, Python, JavaScript, SQL, Bash, SQLite/MySQL, YACC, React, TypeScript
• Libraries & Frameworks: Flask, NumPy, Pandas, OpenCV, APIs, TensorFlow
• Tools: Git, Linux, Unity, Unreal Engine, AWS
COMPUTER SCIENCE EXPERIENCE
Lightcast (Capstone Project) - University of Idaho Project Manager Aug 2024 - Present
• Deployed the website using AWS, enhancing its scalability and reliability
• Developed a tool for managing and automating Excel-based data modeling projects, improving project efficiency and accuracy
• Coordinated between team members to ensure smooth workflow, leading to timely project completion and improved team collaboration
• Conducted regular meetings to track progress and troubleshoot issues, resulting in increased project transparency and problem resolution ELIZA Chatbot - University of Idaho Team Leader Jan 2025 - May 2025
• Built and optimized a rule-based chatbot using Python, regex, and rapidfuzz, improving text understanding and typo handling.
• Enhanced chatbot response precision by 27.5% through iterative improvements in pattern recognition and fallback handling.
• Engineered smarter input normalization, enabling the chatbot to handle real-world misspellings and conversational nuances. HeartbeatAI: ECG Classification using LSTM - University of Idaho Team Leader Jan 2024 - May 2024
• Developed an LSTM-based deep learning model using TensorFlow and Keras to classify ECG heartbeats from the MIT-BIH Arrhythmia Database, achieving an accuracy of 83%.
• Processed data comprehensively, including normalization, reshaping, and splitting (80% training, 20% testing), which optimized model training and improved performance.
• Enhanced model robustness by implementing dropout and batch normalization layers; monitored training efficiency through validation loss and accuracy metrics over 30 epochs.
• Created a graphical User Interface (GUI) for real-time ECG signal prediction visualization, facilitating immediate assessment of model outputs and enabling practical application in healthcare settings.
Monolith (Game Development Project) - University of Idaho Team Leader Aug 2022 - Dec 2022
• Designed and developed 'Monolith,' a game simulating dining services with strategic management challenges, resulting in increased user engagement and positive feedback
• Ensured cross-platform compatibility, enhancing user accessibility and expanding the game's reach across different devices
• Conducted user testing to gather feedback and improve gameplay, leading to a more intuitive user experience and higher satisfaction ratings
• Managed version control using Git for collaborative development, ensuring seamless integration of code changes and reducing conflicts
OTHER EXPERIENCE
Idaho Eatery Chartwells Compass Group SUPERVISOR Aug 2021 - Present
• Managed student employees and scheduled shifts, ensuring operational efficiency and improved team productivity.
• Trained new hires on standard operating procedures, leading to a smoother onboarding process and reduced training time.
• Monitored inventory and placed orders as needed, maintaining optimal stock levels and preventing shortages.
PERSONAL PROJECTS
Al-Powered Mood-Based Playlist Generator
•Engineered an AI-driven web application that detects user mood from facial expressions and dynamically generates personalized Spotify playlists using FastAPI, React, OpenCV, and TensorFlow, enhancing user engagement through intelligent music recommendations.
•Integrated and optimized a full-stack system with asynchronous API calls, Spotify API integration, and a sleek Material UI frontend, ensuring a seamless, real-time user experience with minimal latency.
Google Chrome Dino Game + Neural Networks
•Developed an Al-powered automation system for the Chrome Dino game using NEAT (NeuroEvolution of Augmenting Topologies), enabling the model to evolve and improve its gameplay performance over multiple generations.
•Integrated computer vision (OpenCV, PIL) and reinforcement learning techniques to detect obstacles, capture real-time gameplay, and optimize decision-making for automated gameplay.
EDUCATION
https://github.com/Bibek246
Contact this candidate