- Developed a customized Named Entity Recognition (NER) system to detect company names in news articles. Overcame challenges such as identifying multiple company names in a single sentence and handling company names with similar spellings.- Developed a microservices-based NLP pipeline incorporating several NLP models, including Named Entity Recognition (NER), Topic Modeling and Sentiment Analysis.- Utilized FastAPI to create microservices as part of the pipeline, enabling efficient and… Show more - Developed a customized Named Entity Recognition (NER) system to detect company names in news articles. Overcame challenges such as identifying multiple company names in a single sentence and handling company names with similar spellings.- Developed a microservices-based NLP pipeline incorporating several NLP models, including Named Entity Recognition (NER), Topic Modeling and Sentiment Analysis.- Utilized FastAPI to create microservices as part of the pipeline, enabling efficient and scalable implementation. Designed and implemented a RESTful API for seamless integration with other systems.- Designed the system architecture for the NLP pipeline to ensure scalability and reliability.- Implemented containerization using Docker to package the NLP pipeline into portable and lightweight containers.- Utilized orchestration with Kubernetes to manage and deploy the NLP pipeline across a distributed system, enabling efficient and reliable scaling. Overcame challenges such as load balancing and service discovery. Show less

- Developed a Python script to automate the extraction of ESG information from companies' sustainability reports in PDF format. Utilized OCR (Optical Character Recognition) technology and the Pytesseract library to accurately extract text from PDF documents. The script was able to efficiently extract data from multiple files at once, reducing the time and effort required for manual data extraction.- Conducted extensive data cleaning and pre-processing to prepare the extracted data for… Show more - Developed a Python script to automate the extraction of ESG information from companies' sustainability reports in PDF format. Utilized OCR (Optical Character Recognition) technology and the Pytesseract library to accurately extract text from PDF documents. The script was able to efficiently extract data from multiple files at once, reducing the time and effort required for manual data extraction.- Conducted extensive data cleaning and pre-processing to prepare the extracted data for analysis. This involved removing duplicate and irrelevant information, standardizing the data format, and converting the data into a structured format suitable for machine learning analysis.- Combined rule-based approaches and LDA (Latent Dirichlet Allocation) to generate a comprehensive list of relevant topics. Rule-based approaches involved manually defining a set of rules to identify and extract relevant keywords and phrases related to ESG topics. LDA was used to identify the latent topics present in the data and assign documents to those topics. These techniques were utilized to develop a topic model that accurately identified and classified ESG topics based on the SASB taxonomy.- Achieved 98% accuracy in predicting sentiment using BERT (Bidirectional Encoder Representations from Transformers) for financial reports and news articles related to ESG topics. The BERT model was fine-tuned on a labeled dataset of financial and ESG news articles to accurately classify sentiment as positive, negative, or neutral. By analyzing the language and sentiment used in financial reports and news articles, I was able to develop an algorithm that estimates the overall ESG performance of a company. This algorithm incorporated sentiment analysis scores and financial data to generate an overall ESG score for the company, providing valuable insights into its sustainability practices. Show less

My fiverr profile:https://www.fiverr.com/oussema_guedri

The aim of this project is to develop a platform that connects companies with skilled developers. The platform enables companies to hire the best developers based on their skills and experiences, while providing developers with opportunities to showcase their abilities and increase their visibility in the job market.Tasks/Achievements:- Implement the LinkedIn login functionality using OAuth 2.0 protocol- Create a file upload feature for developers to submit their CV-… Show more The aim of this project is to develop a platform that connects companies with skilled developers. The platform enables companies to hire the best developers based on their skills and experiences, while providing developers with opportunities to showcase their abilities and increase their visibility in the job market.Tasks/Achievements:- Implement the LinkedIn login functionality using OAuth 2.0 protocol- Create a file upload feature for developers to submit their CV- Implement an OAuth-based authentication flow to link a developer's GitHub and LinkedIn accounts- Allow users to update their profile information, such as education and work experience- Build the logic for developers to periodically receive notifications to update their profile- Create a deactivation feature for users to deactivate their account if needed- Develop the backend logic for the HU points Pack System (scoring system)- Create the list of missions and their corresponding backend and frontend- Design a formula for ranking and leveling up, and implement it in the platform- Implement a multi-ranks system for developers to see their rank in different categories- Create a machine learning model to predict a developer's ranking based on their skills and achievements- Allow companies to add custom tests and quizzes- Allow companies to invite external developers to take tests on the platform.- Develop the problem verification feature to ensure tests are valid and reliable- Implement the Judge0 API for checking test results- Build the recommendation system for tests based on a developer's skills and achievements- Scrape and clean data from LinkedIn and other sources to prepare the data for the platform- Augment and enrich the data using natural language processing and other techniques- Create a Flask-based API to serve the platform's functionality Show less

The aim of this project is to develop a cross platform using Bluetooth Low Energy (BLE) technology to establish connections with devices and display collected data. It will also create a user-friendly dashboard to visualize and control the data.Tasks/Achievements:- Research and select a suitable BLE technology and development platform for the application.- Design the user interface and user experience for the application dashboard.- Develop the application's front-end and… Show more The aim of this project is to develop a cross platform using Bluetooth Low Energy (BLE) technology to establish connections with devices and display collected data. It will also create a user-friendly dashboard to visualize and control the data.Tasks/Achievements:- Research and select a suitable BLE technology and development platform for the application.- Design the user interface and user experience for the application dashboard.- Develop the application's front-end and back-end functionalities, including the ability to establish connections with BLE devices and collect data.- Test the application's performance and functionality across different platforms and devices.- Implement security features to ensure the safety of the data collected and transmitted through the application.- Incorporate features that allow for the storage and analysis of data collected from BLE devices.- Develop a user-friendly dashboard that allows for easy visualization and control of data collected from BLE devices.- Incorporate features that allow for real-time notifications and alerts based on the data collected from BLE devices.- Conduct extensive user testing and gather feedback for further improvement and refinement of the application. Show less