AI & Machine Learning Intern @ Samsung Innovation Campus

Faris Mohamed
Kishtah

AI/ML Engineer | RAG & Generative AI Systems | Full-Stack Python | MLOps

Computer & Intelligent Systems student (Future University in Egypt, expected 2027) with hands-on experience building end-to-end RAG, LLM, and MLOps platforms using Python, FastAPI, React, LangChain, Hugging Face, FAISS, ChromaDB, Neo4j, and Docker. Combines applied machine learning — classification, regression, deep learning, computer vision — with production-oriented backend/frontend engineering, observability, and automated testing across generative AI, medical AI, and full-stack software projects.

View Projects Download CV

Cairo, Egypt

Future University in Egypt (Expected 2027)

01 — About Me

Practical AI Engineering &
Production Systems

I am a Computer & Intelligent Systems student at Future University in Egypt (expected graduation 2027), specializing in Retrieval-Augmented Generation (RAG), LLM, machine learning, and MLOps platforms using Python, FastAPI, React, LangChain, Hugging Face, vector databases, and Docker.

My hands-on experience includes Generative AI & RAG engineering at Tips Hindawi, ongoing AI & Machine Learning interning at Samsung Innovation Campus, medical imaging AI, full-stack software architecture, observability, automated testing, and production deployment.

Every project combines applied machine learning — classification, regression, deep learning, computer vision — with production-oriented backend/frontend engineering, model monitoring, test coverage, containerized services, and interactive dashboards.

AI & GenAI Industry Internships

End-to-End AI & MLOps Platforms

BI & Analytics Dashboards Built

English Proficiency (Upper-Intermediate)

02 — Featured Work

Projects & Architectures

★ Featured MLOps Platform

Enterprise AI Manufacturing Platform (Controlled Pilot)

Independently designed and built an end-to-end manufacturing AI/MLOps platform: FastAPI backend, Dramatiq async workers, React/TypeScript/Vite frontend, PostgreSQL, Redis, and a Dockerized observability stack (Prometheus, Grafana, Loki, Alloy). Shipped v1.0.0 with JWT auth and refresh-token rotation, RBAC (admin/engineer/operator), a company-to-factory-to-machine-to-sensor hierarchy, sensor ingestion, and bounded CSV ETL. Built asynchronous training jobs (Random Forest regression/classification) with MLflow tracking and candidate/challenger/champion model governance, plus drift, data-quality, and prediction-audit monitoring, HTTPS deployment, and end-to-end workflow testing.

Python FastAPI React TypeScript PostgreSQL Redis MLflow Prometheus Grafana Docker Drift Detection

GitHub Repo

★ Featured RAG Project

EnterpriseRAG – AI Knowledge Intelligence Platform

Independently designed and built an end-to-end multimodal RAG platform (PDF, DOCX, TXT, audio, video) with a custom hybrid retrieval engine (dense + BM25-style + reranking) and an optional LangChain/LCEL engine using FAISS, PromptTemplate, and PydanticOutputParser. Implemented local LLM inference (Qwen2.5-0.5B-Instruct), embeddings (sentence-transformers/all-MiniLM-L6-v2), and speech-to-text (faster-whisper) behind a FastAPI backend and React/TypeScript frontend. Achieved 6/6 correct answers with 100% retrieval accuracy and citation validity on a deterministic evaluation (60+ backend tests); optimized Compare/Research Report latency 5–8x on an 8GB memory-constrained device.

Python FastAPI React TypeScript SQLAlchemy Alembic Hugging Face Docker Nginx AWS Lightsail

Live Platform GitHub Repo

AI · RAG

AI Document Intelligence Platform

Built a production-style multi-document RAG platform with hybrid BM25/vector retrieval, citation-based answers, Arabic-English support, and Neo4j GraphRAG, using FastAPI, Streamlit, ChromaDB, Docker, and GitHub Actions CI. Implemented document ingestion, chunking, and embeddings alongside agentic report generation, analytics, feedback tracking, retrieval debugging, and confidence indicators.

FastAPI ChromaDB Neo4j BM25 Streamlit Docker CI/CD

Medical AI

NeuroScan AI

Built an AI-assisted brain MRI review application with tumor-pattern prediction, confidence scoring, and Grad-CAM heatmap explanations using deep learning and computer vision, deployed via FastAPI, Docker, and Hugging Face Spaces. Designed an explainable clinical workflow — case history, clinical dashboard, PDF medical-style reports — positioned as a research/decision-support tool.

Deep Learning Grad-CAM FastAPI Docker HF Spaces Computer Vision

Agents & Automation

Self-Improving AI Coding Agent

Built a fully local autonomous coding agent following a Plan → Execute → Test → Reflect → Improve loop, using Pytest feedback and local LLM reasoning via Ollama.

Python Ollama pytest Agentic Loops Local LLMs

Full-Stack Platform

Workflow Orchestration Platform

Built a production-minded admin/operations workflow platform with JWT authentication, RBAC, workflow definitions, approval pipelines, audit logs, and automated API tests using FastAPI, PostgreSQL, Redis, ARQ, and Next.js.

FastAPI PostgreSQL Redis ARQ Next.js JWT & RBAC

ML · Analytics

House Price Prediction App

Modeled Ames Housing data (2,930 records, 82 features) using imputation, encoding, feature selection, and scaling; best model (Random Forest) achieved R² of 0.892 and MAE of approximately $17.8K.

Python Scikit-learn Random Forest Streamlit Pandas

AI · Video

AI-Powered Video Editing MVP

Local AI video editor using Whisper and FFmpeg for automated scene detection, boring-part removal, filler-word removal, subtitle generation, and clip execution.

FastAPI Whisper FFmpeg Python EDL Engine

Agents & Desktop

J.A.R.V.I.S KISHTAH

Desktop AI assistant featuring wake-word interaction, voice responses, application control, screen/camera context awareness, persistent notes, and automation workflows.

Voice AI Local LLM macOS Python Automation

Local AI

Offline Local AI Chat App

Privacy-first offline local AI chat application built with Ollama and Streamlit, featuring RAM/CPU profiling and session memory.

Ollama Streamlit SQLite Profiling

IoT · Embedded

ESP32 Smart Home Security & Safety System

Physical IoT prototype with ultrasonic, gas, flame, and door sensors. Features Firebase real-time database integration and live Streamlit telemetry dashboard.

ESP32 MicroPython Firebase Streamlit

Data Analysis

Business Intelligence & Data Analytics Dashboards

Collection of end-to-end data analytics solutions: Hotel Booking Demand (119K records Excel pipeline), Healthcare Admissions Dashboard (Power BI), and Sales Revenue Analysis (Star Schema).

Power BI Excel Power Query SQL DAX

Robotics & Hardware

ROV Competition — Underwater Vehicle

Remotely Operated Vehicle (ROV) prototype built for competitive underwater operations, featuring custom sensor integration, motor control, and real-time telemetry.

Robotics Electronics Embedded C Sensors

Embedded Systems

6-DOF Robotic Arm & STM32 Smart Garage

Embedded engineering hardware projects including a 6-DOF articulated robotic arm prototype and an STM32-based automated smart garage system.

STM32 Arduino MicroPython Embedded C

03 — Technical Skills

Core Stack & Capabilities

🧠 Programming & Machine Learning

Python C++ SQL NumPy Pandas scikit-learn TensorFlow Deep Learning Computer Vision Model Evaluation

⚡ Generative AI & RAG

LLMs RAG GraphRAG LangChain (LCEL) Hugging Face Transformers FAISS ChromaDB Neo4j Vector Search BM25 Embeddings Prompt Engineering PEFT / LoRA Quantization AI Agents Ollama

⚙️ Backend & MLOps

FastAPI PostgreSQL pgvector SQLite Redis MLflow Docker CI/CD (GitHub Actions) JWT RBAC Model Monitoring Drift Detection Nginx

💻 Frontend & Product

React TypeScript Next.js Streamlit Tailwind CSS Dashboard Design UI / UX

📊 Data & Business Intelligence

SQL Excel Power Query Power BI Pandas NumPy Matplotlib Seaborn EDA KPI Design

🛠️ Testing & Developer Tools

Git GitHub Pytest Playwright

04 — Experience

Work & Internships

AI & Machine Learning Intern

Jul 2026 – Present

Samsung Innovation Campus

Applying data preprocessing, supervised and unsupervised learning, feature engineering, and model evaluation in a structured AI/ML training program using Python, NumPy, Pandas, and scikit-learn.
Translating real-world problems into machine learning tasks and evaluating model performance through Jupyter-based experimentation.

Generative AI & RAG Intern

Jun 2026 – Jul 2026

Tips Hindawi

Completed hands-on training in LLM applications, Hugging Face, Retrieval-Augmented Generation, LangChain, semantic retrieval, vector search, and local model inference.
Built an end-to-end multimodal RAG platform as final project — document extraction, chunking, embeddings, FAISS indexing, LangChain/LCEL chains, PromptTemplate, and PydanticOutputParser for grounded, cited Q&A.
Extended the platform with audio/video transcription and timestamped Q&A, then validated it with backend, integration, and end-to-end tests.

Customer Service Representative

6 months

Concentrix — Amazon Account

Resolved customer inquiries and order-related issues in a high-volume environment while meeting service-quality and performance targets, developing communication, problem-solving, and KPI-driven service skills.

05 — Education & Training

Academic Background & Training

B.Sc. Computer & Intelligent Systems

Expected Graduation: 2027

Future University in Egypt (FUE)

Degree curriculum covering Artificial Intelligence, Machine Learning, Computer Vision, Data Structures & Algorithms, Database Systems, and Intelligent Software Engineering.

Data Analysis & AI Training

Practical Training

Instant Software Solutions

SQL, Excel, Python, Pandas, Power BI, Visualization, Data Cleaning, Web Scraping, Statistics.

Python Programming

Certification

NVIDIA — MCIT Initiative

Specialized Python programming course under the NVIDIA & Ministry of Communications and Information Technology (MCIT) initiative.

06 — Get In Touch

Let's Connect & Collaborate

Open for AI engineering opportunities, RAG system development, and technical collaborations. Feel free to reach out directly.

Email fkishtah@gmail.com

LinkedIn faris-kishtah

GitHub fariskishtah

Phone +20 111 505 5205

Arabic Native

English Upper-Intermediate (B2)

Download Full CV (PDF)

Faris Mohamed Kishtah

Practical AI Engineering &Production Systems