Machine Learning Research Engineer with experience in generative AI, RAG, LLMs, fine-tuning, and MLOps. Has worked on AI chatbots, knowledge retrieval systems, and deep learning applications across robotics, medical devi…
Machine Learning Research Engineer with experience in generative AI, RAG, LLMs, fine-tuning, and MLOps. Has worked on AI chatbots, knowledge retrieval systems, and deep learning applications across robotics, medical devices, and chemical data. Skilled in Python, C++, MATLAB, TensorFlow, PyTorch, Django, Flask, AWS, Docker, and Kubernetes.
TECHNICAL SKILLS
Programming languages C++, Python, MATLAB, JavaScript, Rust, ROS1, ROS2
Core Courses Machine Learning, Computer Vision, Deep Learning, Prompt Engg, Generative AI
AI Language Model Frameworks Huggingface, AutoGPT, AgentGPT, NLTK, Fast.ai, Transformers
CAD Software SOLIDWORKS, CATIA, 3DEXPERIENCE
Deep Learning Frameworks TensorFlow, PyTorch, Keras, SpaCy, Scikit-learn, LightGBM
Vector Databases ChromaDB, FAISS, Weaviate, Qdrant
Programming Libraries Pandas, NumPy, Scikit-learn, CUDA, OpenCV, seaborn, GDAL, SciPy, GPU kernels
Tools PyTest, Jira, Atlassian, Confluence
Web Frameworks Django, Flask, Angular, React
Databases MongonDB, PostgreSQL, MySQL, OLAP, OLTP, SparkSQL
Deep Learning Architecture ResNET, ResNEXT, DenseNET, VGG16, HomographNET, Sfm Learner, U-Net
Platforms Amazon Web Services (AWS), Microsoft Azure ML, Google Cloud Platform (GCP)
Containerization Technologies Docker, Kubernetes
PROFESSIONAL EXPERIENCE
Machine Learning Research Engineer, Co-op at Dassault Syst`emes, Johnston, RI October 2023 - Ongoing
SIMULIA User Assistance AI Chatbot (widgets, Generative AI, RAG, LLMs, Fine-Tuning & Research)
• Pioneered the creation of comprehensive “SIMULIA User Assistance AI Chatbot” for Support Team using
MLOps, integrating advanced Generative AI, RAG (Retrieval-Augmented Generation), Langchain.
• Orchestrated the development of AI Chatbot by seamlessly incorporating automated tone adaptation for engineering
and customer interactions, streamlined Q&A generation with keyword extraction, efficient technical text
summarization, and multi-language translation capabilities.
• Implementation of this chatbot significantly enhanced team efficiency and customer engagement over 95%,
enabling rapid and accurate support in various languages.
• It reduced reliance on the User Success Team, leading to substantial cost savings for the company & eliminating
the need for additional staffing in the teams by 80%.
Dassault System`es Knowledge GPT (CATIA AI Team, RAG, Prompt Engineering, MLOps, Vector Database)
• Led the code Development, Deployment and Testing of Knowledge GPT into the company’s ecosystem using
RAG (Retrieval-Augmented Generation) system, revolutionizing information retrieval across various products by
eliminating the need for extensive manual documentation navigation.
• Architected and deployed a user-friendly GUI utilizing LangChain, Chainlit, and Streamlit, enhancing accessibility
and intuitiveness for users interacting with the knowledge
Pioneered the creation of comprehensive “SIMULIA User Assistance AI Chatbot” for Support Team using MLOps, integrating advanced Generative AI, RAG (Retrieval-Augmented Generation), Langchain.
Orchestrated the development of AI Chatbot by seamlessly incorporating automated tone adaptation for engineering and customer interactions, streamlined Q&A generation with keyword extraction, efficient technical text summarization, and multi-language translation capabilities.
Implementation of this chatbot significantly enhanced team efficiency and customer engagement over 95%.
It reduced reliance on the User Success Team, leading to substantial cost savings for the company and eliminating the need for additional staffing in the teams by 80%.
Led the code Development, Deployment and Testing of Knowledge GPT into the company’s ecosystem using RAG (Retrieval-Augmented Generation) system, revolutionizing information retrieval across various products by eliminating the need for extensive manual documentation navigation.
Architected and deployed a user-friendly GUI utilizing LangChain, Chainlit, and Streamlit, enhancing accessibility and intuitiveness for users interacting with the knowledge base.
Honored by leadership for significantly boosting employee efficiency by 20x providing quick and exact access to a comprehensive mix of images, text, and video documentation.
Engineered and rolled out a 99% precise and efficient user feedback system, to continually refine the LLM for increased accuracy and relevance of responses.
Spearheaded the development of a real-time chemical query chatbot using Bidirectional Encoder Representations from Transformers (BERT) and Large Language Models for rapid response to inquiries about molecular densities, chemical names, formulas, and abbreviations.
Managed a comprehensive chemical database with over 2 million entries, employing APIs and web scraping for data collection, and advanced techniques like normalization and tokenization for data processing.
Leveraged MarkupLM (an LLM variant) for enhanced HTML data extraction, complemented with NLP for proficient query handling.
Innovated and rolled out an LLM-based AI chatbot, reducing chemical data search times from 15 minutes to 10 seconds.
Achieved a 95% adoption rate among engineers, ensuring efficient and precise molecular information retrieval.
Processed tumor ultrasound data, trained a U-Net model for mask prediction, conducted 3D tumor reconstruction, and applied advanced visualization and morphological techniques to the predicted data.
Extracted frames from tumor videos, performed binary segmentation in MATLAB, and trained a U-Net model on the resulting dataset.
Utilized the predicted masks for 3D tumor reconstruction in MATLAB and employed FIJI visualization and morphological operations for refined analysis.
Approach enhanced tumor morphology understanding, potentially improving diagnosis and treatment planning.
Visualization and morphological operations provided comprehensive tumor analysis, supporting future oncology research and clinical applications.
Web Scrapped and collected the data for image facades and labelled them.
Working on Masked Image Modeling to understand architecture age and style through Deep Learning.
نظام التوظيف التشغيلي المعتمد على الذكاء الاصطناعي أولاً، الذي يوحّد ويؤتمت دورة التوظيف بالكامل للمؤسسات الحديثة حول العالم.
د.جوب منطقة حرة ذ.م.م. 2026 © جميع الحقوق محفوظة
