Abdullah Elsaid
AI & Data Engineer specializing in Agentic AI, Large Language Models, NLP, and Data Engineering — turning complex data into autonomous, production-ready AI products.
Engineering Intelligence at Scale
I design AI systems that combine reasoning, retrieval, and robust data engineering to solve real-world problems.
I am an AI & Data Engineer and Computer Science graduate passionate about building systems that can understand, reason, and act. My work sits at the intersection of Generative AI, data engineering, and intelligent automation — where I design agentic workflows, retrieval-augmented generation pipelines, and production-ready ML solutions that create measurable business value.
My journey into AI started with a fascination for how machines can interpret language and make decisions. That curiosity led me to specialize in Artificial Intelligence during my Computer Science studies at Benha University, where I built a strong foundation in algorithms, data structures, databases, and software engineering.
Over the past few years, I have worked across the AI stack — from classical machine learning and deep learning to modern agentic systems powered by LLMs. At Hamber-hub, I developed automation workflows and conversational AI agents using LangGraph, Gemini, and n8n. At NTI and ITI, I sharpened my skills in data analysis, BI architecture, and SQL-driven analytics.
Today, I focus on designing autonomous AI products: systems that can retrieve knowledge, generate insights, interact with databases through natural language, and automate complex workflows with minimal human intervention.
Career Goal
I aim to join a forward-thinking AI team where I can contribute to cutting-edge research and product development — whether that is advancing agentic AI, improving RAG architectures, or scaling LLM-powered applications. My goal is to build technology that is not only intelligent, but also reliable, interpretable, and impactful.
AI-First Mindset
I approach every problem by asking how intelligent systems can automate complexity and augment human decision-making.
Impact-Driven
I focus on measurable outcomes — whether that is faster queries, reduced manual work, or more reliable predictions.
Rapid Learner
From LangGraph agents to BI pipelines, I quickly master new tools and translate them into working products.
Technologies I Work With
A modern AI engineering stack spanning agentic frameworks, machine learning, data engineering, and cloud-ready infrastructure.
AI & LLM Engineering
Machine Learning & Deep Learning
Programming Languages
Backend & APIs
Cloud & DevOps
Databases
Data Engineering & BI
Computer Science Fundamentals
Professional Journey
Hands-on roles across AI engineering, machine learning, and business intelligence — building real-world solutions that create measurable impact.
AI Engineer
Hamber-hub
Designed and deployed AI-powered automation workflows and conversational agents to streamline business operations and improve decision-making velocity.
- Architected AI automation pipelines using n8n to reduce manual data operations and repetitive business tasks.
- Built LLM-powered chatbots with Google Gemini and Agentic AI frameworks for natural, context-aware interactions.
- Implemented Retrieval-Augmented Generation (RAG) to ground chatbot responses in company knowledge and improve accuracy.
- Collaborated with cross-functional teams to integrate AI agents into existing business processes.
Machine Learning for Data Analysis Trainee
National Telecommunication Institute (NTI)
Completed an intensive hands-on program covering the full ML lifecycle from data preprocessing to model evaluation, with real-world projects and case studies.
- Built end-to-end ML solutions using regression, classification, clustering, and ensemble learning techniques.
- Implemented deep learning models including Artificial Neural Networks (ANNs) for predictive analytics.
- Performed data cleaning, exploratory data analysis, feature engineering, and model validation.
Business Intelligence Developer Trainee
Information Technology Institute (ITI)
Completed a 120-hour intensive BI development program focused on data warehousing, advanced SQL, and interactive reporting for data-driven decision making.
- Mastered data warehousing concepts and business intelligence architecture fundamentals.
- Developed advanced SQL skills including query optimization, joins, subqueries, and window functions.
- Built interactive dashboards and reports using Power BI and Tableau to visualize business KPIs.
Featured Work
End-to-end AI projects demonstrating agentic reasoning, retrieval systems, and production-ready machine learning pipelines.
Agentic RAG Chatbot
Intelligent document Q&A with autonomous reasoning
Built an agentic RAG system that dynamically decides whether to retrieve documents, rewrite the user's question, or answer directly — all powered by LangGraph and Google Gemini.
Problem Statement
- Traditional chatbots struggle to answer questions accurately when they need to reason over private documents, refine unclear queries, and decide when to retrieve context versus respond directly.
Architecture
- Document ingestion and embedding storage for semantic search
- LangGraph agent orchestrating retrieval, query rewriting, and response generation
- Google Gemini as the reasoning LLM for context-aware answers
- CopilotKit-powered frontend for conversational interaction
Key Results
- Enabled accurate, grounded answers over user-uploaded documents
- Reduced hallucinations by routing queries to retrieval when confidence is low
- Delivered a modular agent architecture that can be extended with new tools
Lessons Learned
- Agentic routing significantly improves RAG reliability compared to naive retrieval.
- Query rewriting is essential for handling ambiguous or underspecified questions.
Future Improvements
- Add multi-modal document support (images, tables, PDFs)
- Integrate persistent memory for multi-session conversations
- Deploy as a scalable FastAPI service with vector database backend
AI-Powered Inventory Agent
Natural language interface for inventory databases
Developed an autonomous AI agent that translates natural language questions into dynamic SQL queries, executes them against inventory databases, and returns structured insights.
Problem Statement
- Inventory management systems often require technical SQL knowledge, making it difficult for non-technical stakeholders to query stock levels, sales trends, and product information.
Architecture
- LangGraph agent for planning, SQL generation, and result interpretation
- Google Gemini for natural language understanding and query synthesis
- Schema-aware prompt engineering to improve SQL correctness
- Structured data retrieval and analysis from inventory systems
Key Results
- Enabled non-technical users to query inventory data using plain English
- Reduced time-to-insight from manual SQL writing to seconds
- Demonstrated the practical application of agentic AI in enterprise operations
Lessons Learned
- Schema awareness is critical for generating reliable SQL from natural language.
- Agent loops with validation steps reduce query execution errors.
Future Improvements
- Add write-protected safeguards and query approval workflows
- Support multi-table joins and aggregate analytics
- Build a web dashboard for visualizing inventory metrics
NYC Taxi Trip Duration Prediction
End-to-end regression model for spatio-temporal forecasting
Built a supervised machine learning regression pipeline to predict NYC taxi trip durations using spatio-temporal features, robust preprocessing, and model evaluation.
Problem Statement
- Ride-hailing and logistics companies need accurate trip duration estimates to optimize routing, pricing, and customer experience, but real-world data is noisy and geographically complex.
Architecture
- Data preprocessing pipeline handling missing values, outliers, and feature engineering
- Exploratory data analysis and visualization of geospatial and temporal patterns
- Multiple ML algorithms evaluated with MAE and RMSE metrics
- Reproducible training and evaluation workflow using Python data science stack
Key Results
- Achieved strong regression performance with comprehensive feature engineering
- Demonstrated end-to-end ML workflow from raw data to model evaluation
- Applied real-world data cleaning techniques on a large public dataset
Lessons Learned
- Feature engineering on temporal and geospatial data is key to model accuracy.
- Outlier handling and validation strategy directly impact generalization.
Future Improvements
- Experiment with gradient boosting frameworks like XGBoost and LightGBM
- Add interactive map visualization for trip predictions
- Deploy as a FastAPI prediction service
Academic Background
A strong foundation in computer science and artificial intelligence, complemented by intensive industry training programs.
Bachelor's Degree in Computer Science
Artificial Intelligence
Benha University
- Specialized in Artificial Intelligence with coursework in machine learning, deep learning, NLP, and data engineering.
- Built a strong foundation in algorithms, data structures, object-oriented programming, and database systems.
- Completed graduation project on AI-powered inventory management using LangGraph and LLMs.
Certifications
Machine Learning for Data Analysis
National Telecommunication Institute (NTI)
Hands-on training in ML, deep learning, and predictive analytics.
Business Intelligence Developer
Information Technology Institute (ITI)
120-hour program in data warehousing, SQL, Power BI, and Tableau.
Highlights & Milestones
Key accomplishments that reflect technical depth, consistency, and impact across AI engineering and competitive programming.
350+ LeetCode Problems Solved
Sharpened algorithmic thinking and coding efficiency by solving 350+ problems covering data structures, algorithms, and optimization.
View detailsAI-Powered Graduation Project
Developed an autonomous inventory agent as a capstone project, applying LangGraph, LLMs, and NLP-to-SQL in a real-world business context.
View detailsCross-Functional AI Delivery
Collaborated with product and engineering teams at Hamber-hub to ship AI automation workflows and LLM chatbots into production.
Algorithmic Thinking
Consistent practice in competitive programming to sharpen data structure fluency, optimization intuition, and code efficiency.
LeetCode
Competitive Programming
Problems solved and counting
I actively solve algorithmic challenges on LeetCode to strengthen my problem-solving skills, data structure fluency, and ability to write efficient, production-grade code under constraints.
Let's Work Together
Let's build something intelligent together.