Machine Learning Data Science

Isha Patel

Irvine, CA +1-949-***-**** ******@***.*** https://www.linkedin.com/in/ishapatel1108/ https://github.com/Isha1108 EDUCATION

● University of California, Irvine Irvine, CA Sep 2023 - Dec 2024 Master of Computer Science

Coursework: Machine Learning and Generative Models, Artificial Intelligence, Database and Data Management

● K. J. Somaiya Institute of Technology Mumbai, India Aug 2019 - May 2023 Bachelor of Technology - Computer Engineering

Coursework: Software Engineering, Data Structures, Analysis of Algorithms, Operating Systems, Applied Data Science, NLP TECHNICAL SKILLS

Programming Languages: Python, Streamlit, C++/C, Java, R, HTML, CSS, JavaScript, Typescript Frameworks and Libraries: Flask, Selenium, Pandas, Keras, TensorFlow, PyTorch, Sci-Kit Learn, Matplotlib, Seaborn, Beautiful Soup Databases and Tools: MySQL, Git, MS Azure, MS Fabric, MS Office, Power BI PROFESSIONAL EXPERIENCE

Machine Learning Intern Arkision.AI, India Feb 2022 - Jun 2022

● Applied image preprocessing techniques, such as resizing for standardized input dimensions, noise reduction to upgrade image quality, and segmentation to isolate pertinent objects, on real-time surveillance footage.

● Developed an image captioning model achieving 87% accuracy by integrating VGG16 for feature extraction, connecting to LSTM for sequential caption generation, leading to efficient monitoring of the video footage. Python Developer Intern Connect Club, India Jun 2021 - Jul 2021

● Created a robust web application making use of Flask, combined with a MySQL database managed through XAMPP, to provide a startup ecosystem to investors and startups.

● Utilized predictive analytics on historical startup data, empowering investors with insights to make informed decisions and optimize investment strategies based on growth trends. Frontend Web Developer Intern K. J. Somaiya Institute of Technology, India Jun 2020 - Aug 2020

● Designed and crafted a fully functional e-commerce website for a stationery company by leveraging HTML, CSS, and JavaScript.

● Produced custom graphics and fused UX design enhancements using Adobe XD to elevate customer satisfaction and engagement by 20%, resulting in an improved online shopping experience. PROJECTS

ELI - Explain it to me Like I am

● Fine-tuned a Gemini Pro 1.0 large language model (LLM) using Vertex AI to power ELI, specifically enhancing its ability to simplify text across varying comprehension levels.

● Orchestrated the data preparation and training phases, utilizing a comprehensive custom dataset to cover a broad spectrum of text styles and complexities.

● Integrated the fine-tuned LLM into a React-developed Chrome extension, enabling real-time, dynamic text simplification. CancerCare AI Participated in HackTogether 2024, organized by Microsoft Github

● Built a Cancer Care Solution utilizing Microsoft Fabric’s tools: Synapse Data Science, Data Warehouse, and Power BI.

● Deployed a chatbot using Azure OpenAI to boost patient interaction within the Cancer Care AI platform.

● Applied data augmentation methodologies, including CTGAN, to enrich dataset quality and diversity, enhancing the efficacy of machine learning model training and evaluation.

● Managed the database infrastructure adopting Fabric’s Lakehouse architecture, ensuring efficient data storage and retrieval within the Cancer Care Solution ecosystem.

Sparkle Winner at Smart India Hackathon, 2022 GitHub

● Constructed a learning difficulty screening system with 88% accuracy in identifying learning difficulties among children, training an XGBoost model to provide insights into potential types of learning difficulties.

● Leveraged machine learning techniques, including speech recognition and head pose estimation integrated with OpenCV, to innovate the design of screening tests, strengthening accuracy in assessing user responses.

● Elevated learning outcomes and generated detailed progress reports through descriptive data analytics, contributing to a significant improvement in monitoring and lifting each child's intellectual journey. puRED Second Runner up at Irada 2022 Hackathon, organized by L D COE, Ahmedabad GitHub

● Generated a Random Forest classification model with an accuracy of 95% to diagnose PCOS/PCOD based on the user’s symptoms.

● Conceptualized and implemented an intent-based chatbot, using NLP and Tensorflow, to address and alleviate a broad spectrum of menstruation-related queries and concerns.

● Incorporated period cycle monitoring using FullCalendar javascript library to help users track their menstrual health and identify patterns, resulting in an increase in user adoption of these tools. Cure-O-City Finalist in Evathon 2021, organized by University of Limerick, Ireland GitHub

● Pioneered a SVM model for disease prediction based on symptoms, achieving 87% accuracy through model training.

● Employed genetic algorithms for feature selection, systematically optimizing feature subsets, which notably refined the model's accuracy to 94%.

● Seamlessly merged a secure video-calling system through Google Meet API, facilitating convenient remote doctor consultations.

Contact this candidate