Hi, I'm Ali Zamani.
A(n)
Self-driven, quick starter, passionate ML Engineer with a curious mind who enjoys solving complex and challenging real-world problems.
About
🎓 Proud alum of Computer Science at the University of Alberta. 📚
🧩 Enthusiast for problem-solving and coding adventures. 🕵️♂️💻
🚀 Equipped with Python, Tensorflow, GCP, Azure ML, NLTK, MySQL, Power BI, Tableau, Django, and a sprinkle of C++. 🧰✨
👷♂️ 3+ years honing skills in Python, Data Science, and ML sorcery. 💪🔮
💡 Passionate about crafting AI solutions that tackle real-world challenges and positively impact millions. 🌍❤️
Ready to collaborate or chat about tech? Let's connect! 🤝🚀
- Languages: Python, C++, C, MATLAB, PHP, HTML/CSS, JavaScript, SQL
- NLP: NLTK, Spacy, Gensim, Hugging Face, Stanza
- Databases: MySQL, Microsoft SQL Server, SQLite, PostgreSQL
- Libraries: Tensorflow, Pytorch, Keras, Sklearn, Numpy, OpenCV, Scipy, Pandas, React
- Frameworks: Rasa, Laravel, Django, Flask
- Tools: Azure ML Studio, Linux, Git, WordPress, Ns-3 simulator, Docker, NGINX, Bash
- Soft Skills: Communication, Teamwork, Leadership, Work Ethic, Time Management, Creativity
Experience
- Spearheaded the development and support of the Penny Chatbot at Priceline using , Vector RAG, improving customer interaction capabilities.
- Led the creation of a comprehensive ETL pipeline, facilitating a seamless migration from Oracle to Google Cloud Platform, ensuring data integrity and operational continuity.
- Designed and implemented a scalable CI/CD pipeline for Generative AI applications on GCP, specializing in hotel review summarization, significantly enhancing deployment efficiency.
- Tools: Python, GCP, Vertex AI, Git, GitHub, ML pipeline, Teamwork
- Implemented a LightGBM and XGBoost algorithm for predicting the permeability of rock core images with an accuracy of 94%, saving upwards of 10 million dollars for the client.
- Developed a Machine Learning pipeline from scratch on Azure and conducted error analysis to further improve the model performance.
- Conducted error analysis to analyze model performance.
- Prepared bi-weekly update and present it to the client.
- Tools: Python, OpenCV, Azure ML Studio, Git, GitHub, ML pipeline, Teamwork
- Built and implemented the back-end and front-end of the MIRA chatbot.
- Explored and compared different Recurrent Neural Network language models to detect the intent of a sentence and extract entities from it with an F1-score of 97% and 83%.
- Used various data augmentation techniques like back translation and synonym replacement to increase the amount of training data in the MIRA chatbot.
- Applied Sentiment Analysis techniques to MIRA Chabot to identify the sentiment of users’ responses and modify the chatbot’s responses according to detected sentiments.
- Developed a system to automatically report bugs to decrease the time needed for team members to identify and fix bugs/issues.
- Implemented different ways to visualize and send a daily report of MIRA chatbot statistics to team members.
- Trained and managed two undergraduate students in the MIRA chatbot team.
- Developed a service to automatically perform a set of unit tests daily on a product in development to decrease the time needed for team members to identify and fix bugs/issues.
- Tools: Python, Rasa Framework, Tensorflow, Keras, Teamwork
- Experience in leading a group by managing the technical part of CafeIot startup.
- Collaborated with team members utilizing version control systems such as Git to organize modifications and assign tasks.
- Tools: Python, Django, PHP, Laravel, JavaScript, HTML/CSS, Leadership
Projects

An ML pipeline template to create a user-friendly utility to drastically speed up the development and implementation of a machine learning model for all sorts of various problems.

An application on Microsoft Azure for detecting burnout of a call center's agent
- Tools: Microsoft Azure Studio, Python, NLP
- Microsoft and AltaML Hackathon
- Developed an ML pipeline on Azure to detect burnout of a call center’s agent using a pre-trained transformer-based model (BERT).

Determining whether a sentence is sarcastic or non-sarcastic.
- Tools: Python, NLP
- Competed in a Kaggle competitions: Sarcasm Detection with an accuracy of 85+%.

Classifying news as fake and real.
- Tools: Python, NLP
- Competed in two Kaggle competitions: Fake Disaster News Classification, with an accuracy of 90+%.

Identifying sentences that do not make sense and explain why they do not.

Classify a given sentence to one of the four classes of publisher, performer, director, character.
Skills
Languages and Databases











NLP




Libraries









Frameworks




Tools







Visualizations




Soft Skills





Education
Amirkabir University of Technology
Tehran, Iran
Degree: Master of Science in Digital Electronic Systems