Akram Zaytar
Verified Expert in Engineering
AI and ML Developer
Akram is a machine learning engineer specializing in structured and Computer Vision data analysis. After finishing his Ph.D. in applied deep learning, he worked on prototyping and deploying anomaly detection systems. At IBM Research, he guided its intelligence services' climate modeling and ML tooling. Akram excels with end-to-end data-driven intelligence, business requirements analysis, data assembly, and the prototyping, onboarding, and monitoring of AI models.
Portfolio
Experience
Availability
Preferred Environment
PyCharm, MacOS, iTerm2, Postman, Jupyter Notebook, Git, Conda, Python 3, Slack
The most amazing...
...projects I've delivered is a set of educational materials on data science and machine learning that received more than 400 stars on GitHub.
Work Experience
Research Staff Member
IBM Research
- Co-led a research challenge focused on building IBM Environmental Intelligence Suite's CIMF and PAIRS++ toolkits for large-scale climate informatics.
- Built ML-based probabilistic models for sub-seasonal climate forecast post-processing (NN, natural gradient boosting, bayesian methods) by combining different sources of predictability. (Meetingorganizer.copernicus.org/EGU22/EGU22-11063.html.).
- Modeled S2S climate extremes using rule-based and learning methods. This work stream covers both extreme climate featurization and risk estimation.
Postdoctoral Researcher
IBM Research
- Collaborated with KUMWE to provide seasonal maize presence maps in Rwanda. I fused period-based satellite imagery (Sentinel-1 and Sentinel-2) and used ensemble learning methods to solve two machine learning tasks, crop presence and maize potential.
- Investigated the role of extreme weather events on agricultural produce. Used a set of extreme weather indices that quantify extreme weather severity for floods and droughts coupled with XAI algorithms to explain maize yield variation in Iowa.
- Researched ECMWF weather forecasting S5's skill in predicting extreme temperature and precipitation quantile categories ahead of the S2S competition.
Machine Learning Engineer
Tanger-Med
- Built fraud detection systems using weakly-supervised and unsupervised anomaly detection methods based on time-series data.
- Used state-of-the-art deep neural network and machine learning methods in an applied setting using auto-encoders, isolation forests, and LSTMs.
- Deployed the PoC fraud detection system and monitored performance metrics to report on the system's efficacy.
Research Intern
IBM Research
- Presented the work at one of the top-tier machine learning conferences (ICLR). (Cv4gc.org/cv4a2020/.).
- Collaborated with Hello Tractor to use satellite imagery for precision agriculture and small-scale digital farm twinning for IBM Research.
- Deployed a machine learning system for small-scale farm monitoring using large-scale satellite imagery.
Experience
IBM Environmental Intelligence Suite: Probabilistic Weather Forecasting
https://meetingorganizer.copernicus.org/EGU22/EGU22-11063.htmlIBM Environmental Intelligence Suite: Climate Extremes Impact
https://meetingorganizer.copernicus.org/EGU22/EGU22-12461.htmlIBM Intelligence Suite: A Toolkit for Large-scale Geospatial Informatics
https://research.ibm.com/blog/ai-for-climate-change-adaptationEducation
Ph.D. Degree in Applied Artificial intelligence
Faculté des Sciences et Techniques de Tanger - Tangier, Morocco
Master's Degree in Information Systems
Faculté des Sciences et Techniques de Tanger - Tangier, Morocco
Bachelor's Degree in Applied Mathematics and Informatics
Abdelmalek Essaâdi University - Tangier, Morocco
Certifications
Image Processing in Python
Datacamp
Machine Learning with Tree-based Models
Datacamp
Statistical Thinking in Python
Datacamp
Supervised Learning with Scikit-learn
Datacamp
Unsupervised Learning in Python
Datacamp
SQL for Data Science
Datacamp
Cleaning Data in Python
Datacamp
Python Data Science Toolbox
Datacamp
Working with Geospatial Data in Python
Datacamp
Greedy Algorithms, Minimum Spanning Trees, and Dynamic Programming
Coursera
Graph Search, Shortest Paths, and Data Structures
Coursera
Divide and Conquer, Sorting and Searching, and Randomized Algorithms
Coursera
Skills
Libraries/APIs
PyTorch, Scikit-learn, NumPy, Pandas, Matplotlib, SciPy, PyTorch Lightning, OpenCV, Keras, TensorFlow, Shapely
Tools
Jupyter, PyCharm, Sublime Text 3, Postman, Git, Slack
Languages
Python 3, Python, SQL, JavaScript, TypeScript
Paradigms
Data Science, Anomaly Detection
Platforms
Jupyter Notebook, Amazon Web Services (AWS), Linux, MacOS
Storage
Databases
Other
Deep Learning, Learning, Remote Sensing, Artificial Intelligence (AI), Machine Learning, Geospatial Analytics, Neural Networks, AI Programming, iTerm2, Conda, Scientific Research, Data Engineering, Machine Learning Operations (MLOps), Computer Vision, Xarray, Explainable Artificial Intelligence (XAI), Data Visualization, Ensemble Methods, Image Processing, Software Project Management, Big Data, Architecture, Rasterio, Unsupervised Learning, Algorithms, Data Structures, Statistics, Feature Engineering, APIs, Big Data Architecture, Probability Theory, Partial Differential Equations, Measure Theory, Numerical Analysis, Algebra, Software Engineering, Computer Science, System Design, System Administration, Cloud Computing, Research, Satellite Images, Fraud Investigation
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring