Abdullah Ahmad

AI Engineer @ Lean Automation

About

I am an AI Engineer / Machine Learning Engineer / Deep Learning Engineer with a passion for designing and deploying scalable AI solutions that solve real-world problems. My experience spans a range of projects, from building intelligent chatbots to creating deep learning systems for agriculture and trading applications. ► As an AI engineer / AI developer at Lean Automation, I led the development of a smart assessment platform using Hugging Face Transformers and LangChain, enabling dynamic and industry-specific assessment creation with LLMs. I also optimized system performance by integrating retrieval-augmented generation (RAG) and employing advanced data preprocessing techniques using Python, Pandas, and NumPy. ► As a Junior Machine Learning Engineer at Neurog, I developed advanced chatbots using LangChain and RAG, improving response accuracy by 20% and reducing query resolution times. I applied hyperparameter tuning with Optuna, leading to a 15% performance boost, and utilized SQL for efficient data handling while collaborating with teams via Git for version control. ► During my internship at Neurog, I built and evaluated machine learning models such as XGBoost, GRU, and ARIMA, leveraging Scikit-learn for robust performance analysis. By applying feature engineering with Pandas and NumPy, I achieved a 15% increase in the predictive accuracy of a BTCUSDT trading bot. Specialties: - Programming: Python, SQL - Machine Learning & Deep Learning: TensorFlow, PyTorch, Scikit-learn, Keras - Natural Language Processing: Hugging Face Transformers, LangChain, RAG - Model Optimization: Hyperparameter Tuning (Optuna), XGBoost, LightGBM - Data Engineering: Pandas, NumPy, OpenCV - Deployment: Docker, MLOps practices - Version Control: Git With hands-on experience across various domains, I aim to continue leveraging cutting-edge AI technologies to drive innovation and impact. Whether it’s building autonomous systems, optimizing trading strategies, or creating NLP-driven platforms, I thrive in environments where technology meets creativity and problem-solving. Let’s connect and discuss how I can contribute to your team’s success! abdullahrashad3[at]gmail[dot]com

Country

City

Germany

Industry

Computer Software

Skill

n8n, Agentic AI, Written Communication, Research Skills, Writing, Communication, Robotics, MLOps, Google Gemini, DevOps, Project Management, Time Series Forecasting, Generative AI, REST APIs, Flask, IT Operations, GitHub, Git, Natural Language Processing (NLP), Sentiment Analysis

Experience

AI Engineer

Lean Automation

2024-10 - Present · 2 yrs

Led the development of a smart assessment platform using Large Language Models (LLMs) to catalyze digital transformation, resulting in a 30% reduction in manual assessment time. Orchestrated the integration of Retrieval-Augmented Generation (RAG) to dynamically generate tailored assessments, enhancing industry relevance by 25% through proactive problem-solving. Led efficiency enhancements by implementing GraphRAG and segmented vector stores, which independently drove a 20% improvement in retrieval accuracy.

Junior Machine Learning Engineer

Neurog

2024-9 - 2024-10 · 2 mos

Islāmābād, Pakistan

Developed chatbots utilizing Langchain and Retrieval-Augmented Generation (RAG) to streamline user support, enhancing response accuracy and reducing query resolution time. Developed and integrated new chains into chatbot architecture to streamline querying process, leading to a 20% reduction in response time for customer inquiries. Collaborated with other developers to optimize chatbot algorithms for accuracy and efficiency, resulting in a 15% decrease in errors and an overall improvement in user satisfaction.

Machine Learning Intern

Neurog

2024-6 - 2024-9 · 4 mos

Islāmābād, Pakistan

Performed extensive data cleaning and preprocessing on raw OHLC data for trading bot development, removing noise and handling missing values to enhance model reliability and ensure high-quality input. Developed and evaluated multiple machine learning models, including ARIMA, GRU, and Prophet, optimizing performance by refining hyperparameters and assessing accuracy, precision, and recall for real-time predictions in trading bot applications. Utilized feature engineering techniques to enhance the trading bot’s ability to predict BTCUSDT price movements, leading to a 15% improvement in overall trading strategy performance.