Mohammad Razzaghnoori

• Developed and launched four websites using diverse technologies such as Laravel and Vue.js, while collaborating effectively with three different teams. • Optimized over 7 microservices of an application to handle large volumes of data seamlessly, ensuring robust performance and zero failures. • Successfully revitalized and optimized a 5-year-old legacy project initially developed by a junior developer, exceeding client expectations and ensuring satisfaction.

• Worked on Django, and React.js projects side-by-side with the client. • Optimized a client's website for higher conversions and engagement as part of a three-person team. • Created an upsell page that achieved 2,503 conversions within 13 days.

• Developed a closed-domain question-answering system using Python and Flask, integrating with over 10 RESTful domain services for seamless interaction. • Designed and deployed an automatic speech recognition (ASR) data collection bot on Telegram, collecting 200 GB of audio data, which significantly enhanced the accuracy of the ASR system. • Collaborated in a cross-functional team of four, consistently reporting progress and results directly to the CTO and CEO.

• Led the development of the TUTI Platform, a tool for training, testing, and versioning machine learning models, reducing ML tasks' duration by 3 orders of magnitude. • Developed a Named Entity Recognition (NER) model achieving state-of-the-art local F1 score using the multilingual BERT model. • Built and deployed two open-domain question-answering systems: one using various knowledge graphs and the other with advanced information retrieval techniques.