Amit Hirpara

Amit Hirpara #-ï linkedin.com/in/amit-hirpara § github.com/HirparaAmit Summary I possess exceptional proficiency as a practitioner of Machine Learning and Deep Learning, adept at tackling diverse problem statements and uncovering their viable real-world solutions. My expertise lies in crafting models based on intricate theoretical concepts through a meticulous exploration of various approaches, refining algorithms, and persistently striving for optimal results. My ceaseless dedication to honing my skills ensures continuous progress in delivering effective solutions. Education Pandit Deendayal Energy University (PDEU) Bachelor of Technology in Computer Science and Engineering; CGPA: 9.85/10 Ashadeep IIT (Class 12th) GSHSEB - Science; Higher Secondary School Certificate Exam (HSC): 90.00% July 2019 – June 2023 Gandhinagar, Gujarat July 2018 – June 2019 Surat, Gujarat Experience S.S.B. Digital August 2022 – Present Jr. Machine Learning Engineer Ahmedabad, Gujarat • I have diligently undertaken a multitude of Machine Learning (ML) and Deep Learning (DL) endeavors encompassing a wide array of tasks, including sentiment analysis, computer vision tasks, and intricate data extraction tasks, among others. • Leveraging my expertise in TensorFlow and PyTorch, working on the development and implementation of various intricate ML and DL models, starting from the scratch. • Deploying the ML models by creating Flask APIs, thereby facilitating their seamless integration into production environments. Bhaskaracharya National Institute For Space Applications and Geo-Informatics May 2022 – July 2022 Deep Learning Intern Gandhinagar, Gujarat • Worked on the development of a cutting-edge model specifically designed for extracting intricate road and street networks from high-resolution satellite images. • Employing advanced Deep Learning techniques and leveraging the power of Image Segmentation methods, I successfully addressed and resolved the aforementioned problem statement. • For my Semantic Image Segmentation task, I proficiently used the U-Net architecture, a highly effective and widely recognized deep learning model. Projects LipNet | TensorFlow, Deep Learning June 2023 • Created real-time LipNet-based Lip Reading system using TensorFlow for accurate video transcription. • Engineered efficient and scalable Lip Reading pipeline, integrating video preprocessing, feature extraction, and model inference. • Thoroughly tested and evaluated Lip Reading system’s robustness to varying conditions, ensuring reliable performance in diverse scenarios. MIDAS | PyTorch, Deep Learning April 2022 • Created state-of-the-art Monocular Depth Estimation system with PyTorch and OpenCV, enabling accurate depth predictions from live video for diverse computer vision applications. • Implemented state-of-the-art MIDAS architecture, utilizing deep learning, CNNs, and transfer learning for high-quality depth maps in complex visual scenes. Sign Language Detection | TensorFlow, Deep Learning March 2023 • Designed real-time Sign Language Detection system with TensorFlow and OpenCV, identifying key signs: Yes, No, Hello, Thank You, and I Love You. • Developed a custom Jupyter notebook for efficient data collection, annotation, and training of the Sign Language Detection model, ensuring diverse and accurate datasets. • Transfer learning with TensorFlow achieved high precision and recall for specified signs. Implemented efficient bounding box algorithm for sign visualization in output video. Tweet Sentiment Analysis | Natural Language Processing, Machine Learning January 2023 • Developed a highly accurate Tweet Sentiment Analysis system with NLP techniques and classical ML algorithms, achieving 90% accuracy in discerning Positive and Negative sentiments from diverse tweets. • Utilized NLTK library for comprehensive tweet pre-processing, including tokenization, stop-word removal, stemming, and sentiment-specific feature engineering, enhancing ML model input data. • Curated labeled dataset, applied rigorous cross-validation, and hyperparameter tuning for robust, generalized sentiment analysis, accommodating linguistic complexities and informal language use. Image Super Resolution | Generative Adversarial Networks, Deep Learning November 2022 • Designed advanced Image Super Resolution system with TensorFlow, generating high-quality high-resolution images from low-resolution inputs using SRGAN and ESRGAN. • Created user-friendly Flask API for seamless integration of super-resolution model with frontend, enabling real-time image upscaling and user-friendly experience. Skills Languages: Python, C, HTML/CSS, JavaScript, SQL Developer Tools: VS Code, Google Colab, Jupyter Notebook, Google Cloud Platform, PostgreSQL pgAdmin Machine Learning and Deep Learning: TensorFlow, PyTorch, Model Creation, Model Tuning, Model Evaluation, Model Validation Data Science: Explorary Data Analysis, Data Preprocessing, Data Augmentation, Feature Engineering Computer Vision Natural Language Processing Backend/API: Django, Flask, FastAPI, PostgreSQL Publications • Hirpara, A., Patel, S., Vakharia, V., & Kumar, Y. (2023b). A Novel Federated LSTM Model with Conventional LSTM Model for Sentiment Analysis of Twitter Datasets. In International Journal of Advances in Electronics and Computer Science ( IJAECS ), 10(2), 27–36. Certifications Machine Learning with Python: Zero to GBMs | Jovian Programming for Everybody (Python) | Coursera Achievements Intellify - AI Hackathon (2023): Participated in the esteemed artificial intelligence hackathon, hosted by Marwadi University, and cracked the final round, standing out among the 80+ competing teams from across India. Travel Grant (2023): Successfully secured the ”Travel Grant” policy offered by Pandit Deendayal Energy University, enabling me to present my research paper at an esteemed international conference, and representing the university on a global platform. 2nd World Rank in QHack-2022: Achieved the remarkable distinction of being ranked 2nd worldwide in QHack 2022, competing against top talent and industry experts, and contributing to advancements in quantum machine learning technology and its potential applications. Let’s Hack 2.0 (2019): Successfully participated in the prestigious hackathon, organized by Pandit Deendayal Energy University in 2019, and advanced to the final round, securing a place among the top 20 teams out of numerous participants.