Devashri Deulkar

• Collaborated on a Neuromatch Academy 2024 project with the Computational Neuroscience Group Magical Moose.
• Utilized data from Stringer et al. (2019) to decode ocular motion from neural activity in mouse primary visual cortex.
• Built a generalized linear model (GLM) to predict pupil position, change of position, and magnitude
  of movement.
• Project sheds light on sensory processing and neural coding principles in the visual cortex.

• Attending meetings and presentations to learn project details.
• Collaborating with research teams to develop project plans aligned with target timescales.
• Gather, arrange, and corrected research data to create representative graphs and charts highlighting results for presentations.
• Transforming raw data to conform to assumptions of machine learning algorithm.

• MDF(Model Description Format)
  This application simulates and visualizes the SIR model for infectious disease spread.
• Computational graphs, and static, and animated population output are shown at the interface.
  Project: https://huggingface.co/spaces/Devashri/SIR_model
• Demonstrated creativity and resourcefulness through the development of innovative solutions.

• Reproduced dynamics in motor cortical activity using Python and NumPy, demonstrating proficiency in computational neuroscience and data analysis
• Implemented efficient ECoG dataset pre-processing techniques using MNE-python, resulting in improved data filtering and representation for neuroscience research
• Collaborated with team of researchers to analyze and interpret motor cortical activity data, contributing to advancement of knowledge in field of computational neuroscience
• Actively participated in team meetings and discussions, sharing insights and ideas to enhance research methodologies and outcomes

• Implemented "Deep-Emotion: Facial Expression Recognition Using Attentional Convolutional Network" research paper (2019) on facial expression recognition using deep learning techniques, achieving high accuracy and reliability in emotion detection
• Conducted comprehensive training of model using specialized dataset for face emotion recognition, resulting in improved performance and robustness
• Utilized Python, Pytorch, and OpenCV for data preprocessing, model training, and evaluation, demonstrating proficiency in these technologies
• In progress Link to project: https://github.com/Devadeut/Neural-Networks-Hands-On-Projects

• Developed a web-based application for doctor-patient interactions as part of software engineering course project, under guidance of Prof Abir Das at IIT-KGP
• Utilized Python and Django frameworks to build robust and user-friendly interface for seamless doctor-patient interactions. Managed database operations using SQLite3
• Collaborated with team members to design and implement key features, ensuring application met project requirements.
• Conducted thorough testing and debugging to identify and resolve any issues or errors, ensuring application's functionality and reliability.

• Designed and developed an image captioning system using Python and the LAVIS library under the guidance of Prof. Abir Das at IIT-KGP
• Implemented advanced computer vision algorithms to accurately generate captions for images, showcasing proficiency in image processing and natural language processing
• Published the project on GitHub, making it accessible to the wider community and demonstrating a commitment to open-source development and knowledge sharing.

• Engineered coin segregation machine using Arduino UNO and sensor technology, improving the efficiency and accuracy of coin-sorting processes
• Mentored by Prof. M Sinha at IIT-KGP to develop and refine the machine, leveraging expertise in engineering and computer vision.

• Used strong analytical and problem-solving skills to develop effective solutions for challenging situations.
• Resolved problems, improved operations and provided exceptional service.

Grade: CGPA/Percentage: 9.03

Grade: 95.6%, CBSE Board

Grade: 97%, School Topper in CBSE board