Hello, I'm

Ahmed Barakat

Senior AI Engineer

Building intelligent systems at the intersection of AI, Machine Learning, and Automation

Get in Touch View My Work
Scroll to explore

About Me

I'm a Senior AI Engineer with a passion for developing cutting-edge machine learning solutions that solve real-world problems. Currently at TSMC, I'm helping establish and lead a new AI team focused on semiconductor manufacturing optimization.

With a Master's degree in Computer Science from Georgia Tech and a background in Mechatronics Engineering, I bring a unique blend of software expertise and hardware understanding to every project.

My experience spans deep learning, computer vision, NLP, reinforcement learning, and building production-grade ML systems on cloud platforms.

Machine Learning

  • Deep Learning / CNNs / Transformers
  • Computer Vision
  • Natural Language Processing
  • Reinforcement Learning
  • Time Series Analysis

Technologies

  • Python / PyTorch / TensorFlow
  • LangChain / LLMs / GPT
  • Google Cloud Platform
  • BigQuery / Neo4j
  • Docker / Kubernetes

Engineering

  • MLOps & Model Deployment
  • Agentic AI Systems
  • Control Systems
  • Robotics
  • Signal Processing

Work Experience

Senior AI Engineer

TSMC

Phoenix, Arizona Oct 2025 - Present
  • Founding senior engineer of newly established AI team, helping define technical direction, workflows, and initial project roadmap.
  • Collaborated with process and manufacturing teams to develop SPC root cause analysis methods leveraging time-series modeling and Vision-Language Models for chart interpretation.
  • Implemented ESRGAN-based image denoising for Critical Dimension Scanning Electron Microscope (CD-SEM) data to improve measurement quality and increase metrology tool throughput.
  • Mentored junior engineers on visual anomaly detection and wafer defect classification using a pretrained feature extractor backbone model.
  • Leading a small engineering team to build an Agentic AI Virtual Technician Copilot from the ground up, automating complex domain-expert workflows using the LangChain ecosystem.

Automation Engineer Graduate Intern

SIDPEC

Alexandria, Egypt Sep 2024 - Dec 2024
  • Designed and implemented control sequences using control language for Honeywell Distributed Control System (DCS).
  • Configured various points in Honeywell DCS for monitoring, actuation, and control of field instruments and valves. Points included digital composite points for ROV actuation, regulatory control points for general control loops, regulatory PV points for process value calculations like flow compensation, and logic points for process interlocks.
  • Conducted control loop testing and validation for various control loops.
  • Collaborated in the calibration and troubleshooting of instrumentation and valves, including level, flow, and pressure transmitters, as well as remotely operated valves (ROV) and control valves.
  • Identified and troubleshot plant issues by analyzing Piping and Instrumentation Diagrams (P&ID) and wiring diagrams to trace and resolve faults efficiently.

Machine Learning Engineer - Level 2

Paystone

London, Ontario Jan 2023 - Jun 2024
  • Spearheaded the integration of OpenAI's GPT-3.5 API to develop an advanced content generation and completion tool, significantly enhancing efficiency in copywriting processes for the site creation team.
  • Developed a state-of-the-art supervised deep learning model for natural language topic classification of customer reviews, improving insight extraction.
  • Led strategic external negotiations with a data labeling firm, securing top-tier topics labeling services for customer reviews.
  • Designed a comprehensive evaluation suite to meticulously assess the labeling accuracy and performance of vendor services.
  • Implemented and deployed a topics classification service, employing a zero-shot classifier with textual entailment to accurately categorize review content without predefined labels.

Machine Learning Engineer - Level 1

Paystone

London, Ontario Jan 2022 - Dec 2022
  • Engineered and deployed two time series churn prediction models for distinct product platforms, leveraging predictive analytics to enhance customer retention strategies.
  • Successfully operationalized churn prediction models on Google Cloud Platform, utilizing VM Instance Groups for scalable, high-performance production environments.
  • Designed and implemented a robust BigQuery logging middleware to streamline ML request logging, facilitating efficient data analysis and model performance tracking.
  • Collaborated on migrating data from a relational database to a Neo4j graph database to improve customer journey modeling.

Data Intern

Paystone

London, Ontario Apr 2021 - Dec 2021
  • Completed numerous ad hoc data analysis projects, leveraging Machine Learning to extract insights and drive decision-making.
  • Collaborated in a dynamic team environment to develop and deploy interactive PowerBI dashboards.
  • Conducted comprehensive seasonality analyses on client data, identifying key seasonal trends and enabling targeted strategic planning for client engagement.
  • Contributed to a Databricks Proof of Concept (POC), evaluating its utility for enhancing the company's data analytics and processing capabilities.

Projects

Automated Lip Reading

Developed an automated lip-reading model using Densely Connected Temporal Convolutional Network, trained on English and German datasets to decode spoken words from lip movements across languages.

Deep Learning CNN Computer Vision

Machine Translation Transformer

Implemented an Encoder-Decoder self-attention transformer architecture for Many-to-Many machine translation, focusing on German to English translation.

Transformers NLP PyTorch

Multi-Agent Football with PPO

Employed Proximal Policy Optimization with centralized training and decentralized execution to train a multi-agent football team demonstrating advanced coordination strategies.

Reinforcement Learning PPO Multi-Agent

DDQN Lunar Lander

Implemented Double Deep Q-Network architecture in PyTorch to train an agent in OpenAI Gym's Lunar Lander environment, achieving efficient learning and landing strategies.

DDQN PyTorch OpenAI Gym

SLAM Robot Navigation

Developed a Simultaneous Localization and Mapping system to localize a robot in unknown environments, leveraging sensor data and movement controls for real-time mapping.

Robotics SLAM Localization

Kalman Filter Meteor Tracking

Applied Kalman filtering techniques to accurately track meteorites with constant acceleration in 2D simulation based on noisy measurements.

Kalman Filter State Estimation Simulation

Education

Master of Science, Computer Science

Georgia Institute of Technology

Atlanta, Georgia, USA

January 2023 - December 2024

Bachelor of Engineering, Mechatronics Engineering

The University of Western Ontario

London, Ontario, Canada

September 2018 - June 2022

Get In Touch

I'm always interested in hearing about new opportunities, interesting projects, or just connecting with fellow engineers and researchers.

barakat.ahmed.eng@gmail.com
+1 (623) ***-****