I am an aspiring AI Engineer with a strong foundation in Artificial Intelligence, Machine Learning, and Deep Learning. I am skilled in implementing and fine-tuning models like GPT and BERT using Python and open-source libraries. My experience includes optimizing model performance with techniques such as quantization and Mixture of Experts. I am proficient in developing AI applications, including NLP pipelines and classification systems, and I aim to leverage my expertise to build scalable, production-ready solutions.

• 🔭 I’ve worked at Teamlift.co
• 🌱 I’m currently learning Agentic AI, MLOPs, Observability, Scalability
• 👨‍💻 All of my projects are available at yasirrazaa.github.io
• 💬 Ask me about Python, Machine Learning, Deep Learning, LLMs, and MLOps, Agentic AI
• 📫 How to reach me: [email protected]
• 📄 Know about my experiences: yasirrazaa.github.io

• Python
• SQL

• Docker
• Git
• GitHub Actions
• Linux
• Bash
• pandas
• NumPy
• PyTorch
• TensorFlow
• Unsloth
• Transformers
• FastAPI
• Streamlit
• Scikit-learn
• Plotly
• Langchain
• LlamaIndex

• MySQL
• PostgreSQL

• Urdu (Native)
• English (Fluent)

• Collected data from diverse sources using web scraping and API integration.
• Preprocessed and transformed messy web data for further analysis.
• Applied Named Entity Recognition (NER) to extract skills from job descriptions.
• Prepared data for time series forecasting and built models to predict future values.
• Collaborated with data scientists and engineers to design and implement data-driven solutions.

• Built end-to-end pipeline to collect, process, and annotate large amounts of data.

• Coursework: Statistics & Probability, Artificial Intelligence, Databases

• Generated synthetic data using LLMs and validated generated data using Sympy.
• Fine-tuned Numiba-math-7B.

• Built and published an MCP server on PyPi to efficiently work with Jupyter Notebooks when using AI for coding.

• Deployed the SOTA voice agent on Runpod as a serverless endpoint.

• Built a chatbot to chat with documents.

• Implemented a real-time emotion detection using CNN, OpenCV, Streamlit, and WebRTC.

• Built a web app to analyze group and individual chats from a text file.

• Helped a researcher achieve SOTA results on WLASL dataset.

• Deployed the SOTA voice model ChatterBox that beats Eleven Labs on Runpod as a serverless endpoint.

• Advanced LLM Bootcamp: NUST & Sky Electric
• Generative AI Fundamentals, Generative AI, Responsible AI, Intro to Large Language Models: Google
• Data Scientist Track: DataCamp
• Python: Kaggle