Amr Mashlah, Developer in London, United Kingdom
Amr is available for hire
Hire Amr

Amr Mashlah

Verified Expert  in Engineering

Artificial Intelligence (AI) Developer

Location
London, United Kingdom
Toptal Member Since
January 2, 2019

Amr builds machine learning (ML) services from prototype to production. His diverse ML experience helps him build rapid prototypes and find new creative uses for ML models from different domains. Amr enjoys making interactive visualization tools to validate and communicate results.

Portfolio

PatternedAI
Artificial Intelligence (AI)
Eezylife Inc.
Jupyter Notebook, SciPy, Data Visualization, Matplotlib, Scikit-learn...
MachineMedicine
Jupyter Notebook, Data Visualization, Matplotlib, Git, MongoDB, SQLAlchemy...

Experience

Availability

Full-time

Preferred Environment

Jupyter, Git, NumPy, Python, Pandas, Jupyter Notebook, SQL

The most amazing...

...thing I've developed is behavioral clustering with linear discriminant analysis (LDA) using a few labeled data as seeds to influence learned clusters.

Work Experience

Founder

2022 - PRESENT
PatternedAI
  • Developed a customized Stable Diffusion model and served it using serverless GPUs.
  • Fine-tuned a custom image generation model and optimized it for different use cases.
  • Handled autoscaling computes necessary to serve a large volume of users and cut costs with a low number of users.
Technologies: Artificial Intelligence (AI)

Senior Machine Learning Engineer

2020 - PRESENT
Eezylife Inc.
  • Built and maintained recommendation engines for restaurants, events, movies, and music.
  • Extracted key information that helps users relate to their recommendations.
  • Developed an interactive interpretation tool for debugging and validation.
  • Hired and managed a team of data scientists and mentored interns.
Technologies: Jupyter Notebook, SciPy, Data Visualization, Matplotlib, Scikit-learn, Convolutional Neural Networks (CNN), Git, Natural Language Toolkit (NLTK), SQL, SQLAlchemy, Machine Learning, Artificial Intelligence (AI), Keras, Jupyter, NumPy, Pandas, TensorFlow, PostgreSQL, Python, Topic Modeling, Computer Vision

Data Scientist

2018 - 2019
MachineMedicine
  • Used pose estimates from video recording to assess motor skills objectively for Parkinson's patients.
  • Built the analytics pipeline using Python and ingested it in a Flask web application.
  • Created plots to visualize and validate the several steps in the analytics pipeline and the activity detection algorithm.
Technologies: Jupyter Notebook, Data Visualization, Matplotlib, Git, MongoDB, SQLAlchemy, Machine Learning, Artificial Intelligence (AI), Keras, Jupyter, NumPy, Pandas, SciPy, Scikit-learn, Python

Data Scientist

2016 - 2018
IntentHQ
  • Researched new approaches to data enrichment techniques, including behavioral clustering, audience expansion, and modeling user preferences.
  • Enhanced data quality control by creating a web interface for topic disambiguation, automating the repetitive analysis, and reports.
  • Labeled unlabeled data using probabilistic methods.
  • Devised the evaluation metrics for model performance.
Technologies: Jupyter Notebook, Data Visualization, Matplotlib, Git, Natural Language Toolkit (NLTK), SQL, Machine Learning, Artificial Intelligence (AI), Keras, TensorFlow, Jupyter, NumPy, Pandas, SciPy, Scikit-learn, Python

Semantic Search

https://github.com/amrakm/semantic_search
An interactive web app to perform a semantic search in a large number of documents.

A script to embed a list of documents and upload them to a vector database. These embeddings were matched against search queries and served in a Streamlit web app.

Bechdel Test on Movie Scripts

https://github.com/amrakm/BechdalTest
I created a script that scrapes the Internet Movie Script Database (IMSDb) website to apply the Bechdel test to movie scripts.

The script extracts names in the scene, guesses their gender, and runs a test similar to the Bechdel test on each scene—to check if at least two women are talking to each other without the presence of a man in the scene.

ML Framework

https://github.com/amrakm/ML_Framework
A generic ML experiment framework to be used as a starting point and a baseline.

Works on tabular datasets, handles numerical and categorical data automatically, and extracts embedding from text fields using BERT model.

DQN_Navigator

https://github.com/amrakm/DQN_Navigator
Using deep reinforcement learning (DQN) to navigate a 3D Unity ML environment. This is an exercise to train an RL agent using DQN to navigate a large, square world. This is a customized version of Unity ML agents.

Languages

Python, Scala, SQL

Libraries/APIs

Pandas, Keras, NumPy, Scikit-learn, Matplotlib, SQLAlchemy, TensorFlow, SciPy, PyTorch, TensorFlow Deep Learning Library (TFLearn), Natural Language Toolkit (NLTK), Beautiful Soup, Node.js

Tools

Amazon SageMaker, Git, Jupyter

Paradigms

Data Science, Database Design

Platforms

Jupyter Notebook, Amazon Web Services (AWS)

Storage

Databases, PostgreSQL, MySQL, MongoDB

Other

Data Analytics, Machine Learning, Data Cleaning, Data Handling, Machine Language, Convolutional Neural Networks (CNN), Natural Language Processing (NLP), Image Processing, GPT, Generative Pre-trained Transformers (GPT), Large Language Models (LLMs), Artificial Intelligence (AI), Data Reporting, Recommendation Systems, Data Visualization, Dashboards, APIs, Deep Learning, Topic Modeling, Generative Pre-trained Transformer 3 (GPT-3), Sentiment Analysis, Clips, OpenAI, Diffusion Models, Stable Diffusion, DreamBooth, LangChain, Algorithms, Deep Reinforcement Learning, Computer Vision, Vector Data, Vector Databases, Semantics, Scraping, Reinforcement Learning, BERT

Frameworks

Streamlit, Selenium, Next.js

2015 - 2016

Master of Science Degree in Artificial Intelligence

University of Edinburgh - Edinburgh, United Kingdom

2007 - 2013

Bachelor of Engineering Degree in Mechatronics Engineering

University of Aleppo - Aleppo, Syria

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring