- Maintaining and developing the infrastructure enabling the online learning models use in Booking.com
- Distributing training and deployment pipelines for bandits algorithms platform at Booking.com

Machine Learning Engineer

- Experimenting with the use of gMLPs to evaluate the NER models' need for attention modules
- FairScale based distributed training of Charformer-based NER models to handle large-sized input

- Enabling massive-scale deep learning training on distributed systems
- Parallelized models training on multi-GPU clusters

High Performance Computing Engineer

- Optimizing imitation learning strategies for crypto-trading
- Engineering deep reinforcement learning environments for trading agents

- Designed new tailored data augmentation techniques for fracture detection
- Optimized SOTA object detection neural network architectures
- Implemented a genetic ensembling algorithm to improve the inference results

- Automated data collection for user profiles on several online platforms
- Implemented the user-ranking system of the platform
- Wrote a library for Docker containers control and automatic load balancing

Software Engineer

- Conducted a series of predictive analytics for a client canteen to obtain a 23% of food waste decrease
- Events log extraction and process mining visualization for a better client understanding

Data Analyst

- Optimized banking transactions’ ETL queues on a microservices architecture for real-time data analysis
- Built a supervision and performance tracking dashboard for the backend services

- Centralized an e-commerce online sales control through an ERP and performance monitoring dashboard

Along with studies, I participated & won a lot of data science hackathons around deep learning & artificial intelligence in general. I also used these technologies to apply the deep learning concepts on different tasks including computer vision & object detection, finance & crypto trading, and even NLP & language models.

Personal details

Ramy K.

Summary

Technical skills

Work Experience